Intelligent negotiation and reservation method for sports practice teaching training course time

Abstract

The application discloses an intelligent negotiation and reservation method for sports practice teaching training course time, comprising the steps of establishing a virtual agent with a preference learning ability for students, coaches and fields in the system, introducing an adaptive negotiation mechanism, effectively solving the technical problems of a traditional sports training course reservation system in multi-agent dynamic balance, dynamic environment resistance, multi-constraint conflict processing and notification distribution control, providing a more intelligent, efficient and robust sports training course time negotiation and reservation method, and significantly improving the intelligent level of course arrangement and scheduling and user experience.

Description

Technical Field

[0001] This invention belongs to the field of intelligent scheduling technology, specifically relating to an intelligent negotiation and reservation method for the time of physical education practical teaching and training courses. Background Technology

[0002] With the popularization of national fitness awareness and the rapid development of the sports training industry, the scheduling of classes at sports venues and training institutions (or sports practical teaching and training) faces increasingly complex challenges. Traditional sports practical teaching and training course reservation systems mostly adopt a centralized, rule-driven, hard-matching logic, meaning the system allocates courses to students based on a static coach schedule and venue availability, following a first-come, first-served principle. However, this traditional method has revealed many technical shortcomings in practical applications:

[0003] First, traditional systems lack the ability to dynamically balance the demands of multiple stakeholders. Class scheduling is essentially a game of interests among students' time preferences, instructors' physical condition, and the availability of venue resources. Existing technologies typically only meet the minimum requirement of "no time conflicts," neglecting students' preferences, the need for buffering and resetting venue facilities, and the accumulated physical fatigue of instructors after continuous high-intensity teaching. This easily leads to scheduling plans being significantly compromised during implementation due to instructors' exhaustion or the venue not being cleaned in a timely manner.

[0004] Secondly, existing scheduling algorithms cannot effectively withstand the disturbances of dynamic environmental resistance. For example, sudden factors such as traffic congestion and extreme weather can significantly reduce the fulfillment rate of instructors or students. Traditional algorithms often lack a mechanism to convert these non-time-dimensional physical environmental parameters into scheduling constraints, resulting in extremely low fulfillment rates of the generated timetables under adverse conditions.

[0005] Furthermore, traditional heuristic search or linear programming scheduling algorithms often use simple linear weight addition to calculate priorities when dealing with multiple constraint conflicts. This linear model not only faces the barrier of difficulty in unifying different physical dimensions, but also is prone to logical loopholes such as positive and negative factors canceling each other out under boundary conditions (i.e., the "factor failure" phenomenon), causing the system to fail to truly reflect the absolute veto power of extreme fatigue or extreme will on the scheduling results. At the same time, when the system encounters scheduling conflicts, it often blindly pushes massive amounts of coordination notifications to various terminals, resulting in wasted channel resources and information overload. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent negotiation and reservation method for sports practice teaching and training course times, in order to solve the problems mentioned in the background art. The intelligent negotiation and reservation method for sports practice teaching and training course times provided by this invention has the following characteristics.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for intelligent negotiation and reservation of sports practice teaching and training course time, comprising establishing virtual agents with preferred learning abilities for students, coaches, and venues within the system, and further comprising the following steps:

[0008] S1. Extract the equivalent time consumption of multidimensional historical behavioral features and real-time dynamic environmental resistance, and quantify and derive the student's willingness preference score, coach's time preference score, real-time performance prediction rate and venue expected turnover rate to form the multidimensional preference and state parameters of each virtual agent.

[0009] S2. In response to the course reservation request, the target time is discretized, and a hard-constrained joint access matrix is ​​constructed by combining the reference value of the site's intrinsic facility reset time and the reference value of the coach's intrinsic physical recovery time. Candidate time slot combinations are filtered and generated, and the initial negotiation utility expectation of each virtual agent for the candidate time slot combinations is initialized.

[0010] S3. Trigger multiple rounds of dynamic negotiation. Each virtual agent executes a concession strategy based on its own multidimensional preferences and state parameters. Calculate the dynamic concession tolerance of each virtual agent and adjust and iteratively update the expected utility value of each virtual agent in this round accordingly.

[0011] S4. Couple the expected value of each virtual agent's iteration utility in the current round in multiple dimensions, calculate the global comprehensive negotiation utility index, and determine the negotiation convergence when the index reaches the preset global utility threshold, and take the current candidate time period combination as the final feasible solution.

[0012] S5. Generate the final training course schedule based on the feasible solution, and reverse map the final global comprehensive negotiation utility index output when the negotiation converges into the final target notification distribution amount, and distribute the scheduling instructions to the corresponding terminals according to the final target notification distribution amount.

[0013] In this invention, in S1, the multidimensional historical behavioral features include the total number of actual attendances of students within a set period extracted from historical logs, the historical performance benchmark rate of coaches and the historical average check-in delay time, and the current cumulative load duration of the venue; the real-time dynamic environmental resistance equivalent time includes the real-time traffic congestion equivalent time and extreme weather resistance equivalent time obtained through external interfaces, and the sum of the two constitutes the real-time dynamic environmental resistance equivalent time.

[0014] In this invention, further, in S1, the method for calculating the equivalent time consumption of real-time dynamic environmental resistance includes the following steps:

[0015] S111, Calculate the equivalent time of real-time traffic congestion:

[0016] ;

[0017] in, This indicates the equivalent time spent in real-time traffic congestion. This indicates the expected arrival time under the current real-time traffic conditions. Indicates the baseline driving time. This represents the penalty coefficient for unexpected road blockage;

[0018] S112. Calculate the equivalent time of drag during extreme weather:

[0019] ;

[0020] in, This represents the equivalent time spent due to drag during extreme weather events. Based on the baseline driving time, Indicates the travel sensitivity coefficient. This indicates real-time precipitation or wind intensity index. Indicates the circuit breaker threshold for extreme weather events;

[0021] S113. Calculate the equivalent time consumption of real-time dynamic environmental resistance:

[0022] ;

[0023] in, This represents the final output, which is the equivalent time taken to generate real-time dynamic environmental resistance. This indicates the equivalent time spent in real-time traffic congestion. This indicates the equivalent time taken for resistance during extreme weather events.

[0024] In this invention, further, in S1, the method for quantifying and deriving multidimensional preferences and state parameters by utilizing multidimensional historical behavioral characteristics and the equivalent time consumption of real-time dynamic environmental resistance includes the following steps:

[0025] S121. Calculate the score of trainee preference:

[0026] ;

[0027] in, This represents the score indicating the student's willingness and preference. This represents the total number of actual attendances in history. This indicates the threshold number of times a basic habit has been formed;

[0028] S122. Calculate the coach's time preference score:

[0029] ;

[0030] in, This indicates the coach's time preference for scoring. This indicates the amount of free time available for the coach. Indicates the standard duration of a single course;

[0031] S123. Calculate the real-time fulfillment prediction rate:

[0032] , ;

[0033] in, Indicates the real-time fulfillment prediction rate. Indicates the historical performance benchmark rate. This represents the equivalent time consumed by real-time dynamic environmental resistance. Indicates the time consumed by the tolerance boundary of extreme environmental resistance. This indicates the total number of times a coach has actually fulfilled their contractual obligations within a set period. This indicates the total number of fixed training sessions a coach has scheduled within the same period.

[0034] S124. Calculate the expected site turnover rate: ;

[0035] in, This indicates the expected turnover rate of the venue. This indicates the current cumulative load duration. This indicates a reference value for the physical maintenance cycle of the site's intrinsic facilities.

[0036] In this invention, in S2, the reference value for the intrinsic facility reset time of the venue is the minimum buffer time required between two uses of the venue; the reference value for the intrinsic physical recovery time of the coach is the minimum recovery time required after the coach has taught continuously.

[0037] In this invention, further, in S2, the method for constructing the hard-constrained joint passage matrix and filtering out the initial candidate time period combinations includes the following steps:

[0038] S211, Discretize the target reservation time period into A time slice, for a time slice Determine the characteristic elements of the open state of the venue and the characteristic elements of the coach's teaching boundary state respectively;

[0039] The logical determination model for the characteristic elements of the site's open status is as follows:

[0040] ;

[0041] in, Elements indicating the open status of a site Indicates time slice Available idle time before and after, This indicates a reference value for the time required to reset the site's intrinsic facilities;

[0042] The logical judgment model for the boundary state feature elements of coaching instruction is as follows:

[0043] ;

[0044] in, Represents the feature elements of the teaching boundary state. This indicates the cumulative continuous teaching time of the coach corresponding to the allocated time slot. This indicates the reference value for the coach's intrinsic physical recovery time. Indicates the safety factor for extreme teaching conditions;

[0045] S212, will and Perform element-wise multiplication to construct the initial candidate time period mask vector:

[0046] Filter mask vector The time slices with a product of 0, and the consecutive time slices with a product of 1 and a corresponding cumulative duration greater than or equal to the standard single course duration, are used as the initial candidate time slot combination.

[0047] In this invention, further, in S2, the method for evaluating and calculating the initial negotiation utility expectation of each virtual agent for the candidate time period combination includes the following steps:

[0048] S221. Calculate the initial negotiation utility expectation of the student's virtual agent:

[0049] ;

[0050] in, This represents the initial negotiation utility expectation of the student's virtual agent. This represents the score indicating the student's willingness and preference. This represents the time offset between the candidate time period and the most desired target time period. This indicates the student's maximum tolerance for time offset.

[0051] S222. Calculate the initial negotiation utility expectation of the coach virtual agent:

[0052] ;

[0053] in, This represents the initial negotiation utility expectation of the coach's virtual agent. This represents the preset weighting coefficient. This indicates the coach's time preference for scoring. This represents the historical average check-in delay time extracted from historical behavioral characteristics. This indicates a reference value for the standard course buffer time;

[0054] S223, Expected initial negotiation utility of the virtual agent for the computational site:

[0055] ;

[0056] in, This represents the initial negotiation utility expectation of the virtual agent for the venue. This indicates the expected turnover rate of the venue. This indicates the available time interval between the candidate time slot and the existing schedule adjacent to the venue. This indicates a reference value for the time required to reset the site's intrinsic facilities.

[0057] In this invention, further, in S3, the process of calculating the dynamic concession tolerance of each virtual agent and correcting the expected value of the iterative update utility includes the following steps:

[0058] S31, in the During the negotiation rounds, for the virtual coach agent, the coach's dynamic concession tolerance is calculated by combining its cumulative continuous teaching hours with the real-time performance prediction rate, and its iterative utility value is updated accordingly.

[0059] , ;

[0060] in, This indicates the coach's dynamic tolerance for concessions. This indicates the coach's tolerance for basic concessions. This indicates the coach's cumulative continuous teaching hours. This indicates the reference value for the coach's intrinsic physical recovery time. Indicates the real-time fulfillment prediction rate. Indicates the historical performance benchmark rate. Indicates the coach virtual agent number Round iteration utility value, This represents the initial negotiation utility expectation of the coach's virtual agent;

[0061] S32. For the site virtual agent, calculate the site's dynamic concession tolerance based on the site's current cumulative load duration, and update its iterative utility value:

[0062] , ;

[0063] in, Indicates the tolerance for dynamic concessions on the field. Indicates the tolerance for site foundation concessions. This indicates the current cumulative load duration. This indicates a reference value for the physical maintenance cycle of the site's intrinsic facilities. Indicates the current round of negotiation interaction. This indicates the maximum number of negotiation rounds. Indicates the venue virtual agent number Round iteration utility value, This represents the initial negotiation utility expectation of the virtual agent for the venue;

[0064] S33. For virtual agents of students, introduce a willingness-to-preference integral to calculate the student's dynamic concession tolerance and update its iterative utility value:

[0065] , ;

[0066] in, Tolerance should be reduced to accommodate student dynamics. This indicates the student's basic tolerance level for concessions. This represents the score indicating the student's willingness and preference. Indicates the current round of negotiation interaction. This indicates the maximum number of negotiation rounds. Indicates the number of virtual agents for students. Round iteration utility value, This represents the initial negotiation utility expectation of the student's virtual agent.

[0067] In this invention, further, in S4, the formula for calculating the global comprehensive negotiation utility index is as follows:

[0068] ;

[0069] in, Indicates the first The overall comprehensive negotiation utility index of this candidate time period combination in the round of negotiation. , , These represent the first, second, and third virtual agents of students, coaches, and venues, respectively. Round iteration utility value.

[0070] In this invention, further, in S5, the control model for inversely mapping the final global comprehensive negotiation utility index into the final target notification distribution amount is as follows:

[0071] When the global comprehensive negotiation utility index reaches a preset threshold and the iteration terminates, and this threshold is determined as the final global comprehensive negotiation utility index, the following control calculations are performed:

[0072] ;

[0073] in, This represents the final target notification distribution amount generated by the system to control the scheduling of underlying channel resources. This represents the system's preset baseline distribution frequency value. This represents the final, comprehensive, and negotiated utility indicator.

[0074] Compared with the prior art, the beneficial effects of the present invention are:

[0075] 1. By introducing an adaptive negotiation mechanism, this invention effectively solves the technical problems of traditional sports training course reservation systems in terms of dynamic balance of multiple subject demands, resistance to dynamic environmental resistance, handling of multiple constraint conflicts, and notification distribution control. It provides a more intelligent, efficient, and robust method for negotiating and reserving sports training course times, significantly improving the intelligence level of scheduling and user experience.

[0076] 2. The construction of multidimensional preferences and state parameters of the virtual agent in this invention is more accurate and dynamic. These precise parameters, as the basis for negotiation, can more realistically reflect the actual situation of students, coaches, and venues, as well as the potential risks of the external environment. This enables multi-round dynamic negotiation to be based on more accurate information, and the final calculated global comprehensive negotiation utility index is more valuable. This ensures that the final training schedule not only meets the basic needs of all parties, but also maximizes the satisfaction and fulfillment of all parties, significantly improving the accuracy, reliability, and adaptability of the intelligent negotiation and appointment method.

[0077] 3. This invention can effectively transform raw data such as the total number of students' historical actual attendance, the available free interval of coaches, the historical performance benchmark rate, the equivalent time of real-time dynamic environmental resistance, and the current cumulative load duration of the venue into student willingness preference scores, coach time preference scores, real-time performance prediction rates, and venue expected turnover rates with clear quantitative significance. These quantified multidimensional preference and state parameters provide a unified and comparable decision basis for each virtual agent in the system, enabling virtual agents to more accurately assess their own acceptance, preference, and performance risks for different course periods.

[0078] 4. In the early stages of course booking requests, this invention can efficiently eliminate a large number of time slots that do not meet the basic conditions based on the objective hard constraints of venue and instructor. This pre-filtering mechanism greatly reduces the search space for subsequent dynamic negotiation, avoids complex utility evaluation and concession strategy calculation for unreasonable time slots, and thus significantly improves the efficiency and accuracy of the entire intelligent negotiation and booking method. It ensures that the candidate time slot combinations entering the negotiation stage are physically and physiologically feasible, providing a solid foundation for subsequent refined preference matching and utility optimization, so that the final generated timetable not only meets the preferences of all parties, but also conforms to the feasibility of actual operation.

[0079] 5. This invention effectively solves the problem of lacking objective and quantitative initial evaluation of the virtual agents of each party for the candidate time slot combination in the intelligent negotiation and reservation method. After generating the candidate time slot combination, it provides a solid and multi-dimensional evaluation benchmark for the subsequent dynamic negotiation process by accurately calculating the initial negotiation utility expectations of the student virtual agent, coach virtual agent and venue virtual agent. This makes the negotiation no longer a blind trial, but a quantitative starting point based on the real preferences and constraints of each party.

[0080] 6. The initial utility expectation of the trainees in this invention takes into account their learning willingness and time flexibility, the initial utility expectation of the coaches takes into account their time preference and contract fulfillment reliability, and the initial utility expectation of the venue balances its turnover efficiency and facility maintenance needs. This comprehensive initial assessment can significantly improve the efficiency and success rate of negotiation, avoid negotiation deadlocks or inefficient cycles caused by information asymmetry or unclear initial expectations. Finally, by providing candidate solutions that better meet the initial wishes of all parties, the system can be prompted to converge to a scheduling result acceptable to all parties more quickly, thereby improving the overall user satisfaction and system operation efficiency of intelligent negotiation and appointment of sports training courses.

[0081] 7. This invention can effectively solve the problems of rigidity and stagnation that may occur in the process of multi-agent negotiation. During the negotiation process, each virtual agent no longer just adheres to its initial preferences, but can dynamically adjust its willingness to make concessions according to its own real-time status (such as the coach's fatigue level, the load of the venue) and the progress of the negotiation (such as the negotiation round). This mechanism makes the negotiation process more flexible and can simulate the compromise behavior of participants in the real world when facing conflicts, thereby significantly improving the efficiency and success rate of negotiation.

[0082] 8. This invention effectively solves the problem of comprehensively evaluating the utility values ​​of each virtual agent in multi-party negotiation. Through the coupling method of geometric average, it ensures that the global comprehensive negotiation utility index can accurately reflect the balance of interests of all parties, avoiding excessive unfairness to any party in the negotiation result. This enables the system to more accurately judge the timing of negotiation convergence and select the final scheduling scheme that is highly acceptable to all participants, thereby significantly improving the fairness, stability and efficiency of intelligent negotiation and reservation of sports training course time.

[0083] 9. This invention reverse-maps the final global comprehensive negotiation utility index into the final target notification distribution volume. This solution can intelligently adjust the notification strategy of scheduling instructions according to the quality of the negotiation results. When the utility index of the negotiation results is high, it indicates that all parties are highly satisfied with the time period, and the system can notify at an appropriate frequency to avoid wasting resources. When the utility index is low, the system will increase the notification distribution volume and remind relevant parties in a more frequent or urgent manner, thereby effectively reducing the performance risk caused by unsatisfactory negotiation results and ensuring the smooth progress of the training course. Attached Figure Description

[0084] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0085] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0086] In the description of this invention, unless otherwise explicitly specified and limited, the terms "connected," "linked," and "fixed" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0087] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0088] In the description of this invention, the terms "upper," "lower," "right," and "left," etc., refer to the orientation or positional relationship shown in the accompanying drawings. They are used only for ease of description and simplification of operation, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention. Furthermore, the terms "first" and "second" are used merely for distinction in description and have no special meaning.

[0089] Example 1

[0090] Please see Figure 1 This embodiment provides the following technical solution: a method for intelligent negotiation and reservation of sports practice teaching and training course time, including establishing virtual agents with preference learning capabilities for students, coaches and venues within the system, and also including the following steps:

[0091] S1. Extract the equivalent time consumption of multidimensional historical behavioral features and real-time dynamic environmental resistance, and quantify and derive the student's willingness preference score, coach's time preference score, real-time performance prediction rate and venue expected turnover rate to form the multidimensional preference and state parameters of each virtual agent.

[0092] S2. In response to the course reservation request, the target time is discretized. The hard-constrained joint access matrix is ​​constructed by combining the reference value of the venue's intrinsic facility reset time (the minimum buffer time required between two uses of the venue (in minutes)) and the reference value of the coach's intrinsic physical recovery time (the minimum recovery time required after continuous teaching by the coach (in hours)). Candidate time slot combinations are filtered and generated. The initial negotiation utility expectation of each virtual agent for the candidate time slot combinations is initialized.

[0093] S3. Trigger multiple rounds of dynamic negotiation. Each virtual agent executes a concession strategy based on its own multidimensional preferences and state parameters. Calculate the dynamic concession tolerance of each virtual agent and adjust and iteratively update the expected utility value of each virtual agent in this round accordingly.

[0094] S4. Couple the expected value of each virtual agent's iteration utility in the current round in multiple dimensions, calculate the global comprehensive negotiation utility index, and determine the negotiation convergence when the index reaches the preset global utility threshold, and take the current candidate time period combination as the final feasible solution.

[0095] S5. Generate the final training course schedule based on the feasible solution, and reverse map the final global comprehensive negotiation utility index output when the negotiation converges into the final target notification distribution amount, and distribute the scheduling instructions to the corresponding terminals according to the final target notification distribution amount.

[0096] The following example will provide a more detailed explanation of the above technical solution:

[0097] Suppose user A wants to book a sports training class, preferring a Wednesday afternoon from 2 PM to 4 PM. The system first establishes virtual agents for user A, coach B, and venue C. These virtual agents analyze user A's historical attendance records, coach B's historical teaching duration and contract fulfillment, and venue C's historical usage and maintenance records to initially learn and store their respective preference information. For example, user A's agent might learn that user A has a higher preference for the afternoon time slot, coach B's agent might learn that coach B needs a break after teaching for 4 consecutive hours, and venue C's agent might learn that venue C requires 30 minutes of cleaning time after each use. Next, the system extracts multidimensional historical behavioral features and the equivalent time consumption of real-time dynamic environmental resistance. For example, the system obtains from historical logs the total number of actual attendances of user A in the past month, coach B's historical performance benchmark rate and average check-in delay time, and the current cumulative load duration of venue C. Simultaneously, the system obtains current real-time traffic congestion and weather forecasts through external interfaces, calculating the equivalent time consumption of real-time traffic congestion and extreme weather resistance. This data is used to quantitatively derive user A's willingness preference score, coach B's time preference score, real-time performance prediction rate, and venue C's expected turnover rate, thus constituting the multidimensional preferences and state parameters of each virtual agent. For example, if real-time traffic congestion is severe, coach B's real-time performance prediction rate may decrease. When user A submits a class booking request for Wednesday afternoon from 2 PM to 4 PM, the system responds and discretizes the target time into multiple time slices, for example, discretizing 2 PM to 4 PM into four and a half hour time slices. The system combines the intrinsic facility reset time reference value of venue C (e.g., 30 minutes) with the intrinsic physical recovery time reference value of instructor B (e.g., 2 hours) to construct a hard-constraint joint access matrix. This matrix filters out those classes between 2 PM and 4 PM if venue C just finished its previous class at 1:30 PM, or if instructor B has already taught continuously for more than [number missing] hours before 2 PM. The system generates a series of candidate time slot combinations that meet basic physical and physiological constraints through a 3-hour time slice. For example, if Wednesday 2 PM to 4 PM does not meet the conditions, the system may generate Wednesday 4 PM to 6 PM or Thursday 2 PM to 4 PM as candidate time slot combinations. Subsequently, each virtual agent initializes the initial negotiated utility expectation for these candidate time slot combinations based on its current multidimensional preferences and state parameters. For example, user A's agent may consider Wednesday 2 PM to 4 PM to have the highest utility expectation, while Thursday afternoon's utility expectation is slightly lower.Subsequently, the system triggers multiple rounds of dynamic negotiation. In each round, the virtual agents of User A, Coach B, and Venue C execute concession strategies based on their respective multidimensional preferences and state parameters. For example, if the current negotiation candidate time slot is Wednesday afternoon from 4 PM to 6 PM, User A's agent might calculate its dynamic concession tolerance. Considering User A's overall preference for the afternoon time slot, its concession tolerance might be relatively low. Coach B's agent will calculate its dynamic concession tolerance based on Coach B's cumulative continuous teaching time and real-time performance prediction rate. If Coach B has already taught a lot that day and its real-time performance prediction rate is low, its concession tolerance will also be low. Venue C's agent will calculate its dynamic concession tolerance based on Venue C's current cumulative load duration. If Venue C's usage rate is high that day, its concession tolerance might be high. Based on these dynamic concession tolerances, each virtual agent will adjust and iteratively update the expected utility value for this round. For example, if User A's concession tolerance is low, the decrease in its expected utility value will be smaller, and vice versa. After each round of negotiation, the system couples the expected utility values ​​of the virtual agents of user A, coach B, and venue C in the current round in multiple dimensions to calculate a global comprehensive negotiation utility index. For example, the system may use a geometric mean to couple the expected utility values ​​of the three agents to ensure that if the expected utility of any one party is too low, it will significantly lower the overall index. When the global comprehensive negotiation utility index reaches a preset global utility threshold (e.g., 0.8), the system determines that the negotiation has converged and takes the current candidate time period combination (e.g., Wednesday afternoon from 4 pm to 6 pm) as the final feasible solution. Finally, the system generates the final training schedule based on the feasible solution. At the same time, the system will reverse-map the final global comprehensive negotiation utility index (e.g., 0.85) output at the negotiation convergence to the final target notification distribution volume. For example, if the global utility index is high, the system may calculate a lower notification distribution volume and send scheduling instructions only to user A, coach B, and venue C. If the global utility index is low, the system may calculate a higher notification distribution volume and, in addition to sending scheduling instructions, may also send warning notifications to relevant management personnel for manual intervention or subsequent optimization. Based on the final target notification distribution volume, the system distributes scheduling instructions to the corresponding terminals to complete the course reservation.

[0098] By adopting the above technical solution, this invention effectively solves the technical problems of traditional sports training course reservation systems in terms of dynamic balance of multiple subject demands, resistance to dynamic environmental resistance, handling of multiple constraint conflicts, and notification distribution control by introducing an adaptive negotiation mechanism. It provides a more intelligent, efficient, and robust method for negotiating and reserving sports training course times, significantly improving the intelligence level of scheduling and user experience.

[0099] Specifically, the data structure for the equivalent time of multidimensional historical behavioral characteristics and real-time dynamic environmental resistance is defined as follows: Multidimensional historical behavioral characteristics include the total number of actual attendances of students within a set period extracted from historical logs, the historical performance benchmark rate of coaches, the historical average check-in delay time, and the current cumulative load duration of the venue; Real-time dynamic environmental resistance equivalent time includes the equivalent time of real-time traffic congestion and the equivalent time of extreme weather resistance obtained through external interfaces, and the sum of the two constitutes the equivalent time of real-time dynamic environmental resistance.

[0100] The specific steps for calculating the equivalent time consumption of real-time dynamic environmental resistance are as follows:

[0101] S111, Calculate the equivalent time of real-time traffic congestion:

[0102] ;

[0103] in, This indicates the equivalent time of real-time traffic congestion (in minutes). This indicates the estimated arrival time (in minutes) under the current real-time traffic conditions. This indicates the baseline driving time (in minutes). This represents the penalty coefficient for unexpected road blockage;

[0104] S112. Calculate the equivalent time of drag during extreme weather:

[0105] ;

[0106] in, This represents the equivalent time (in minutes) of drag during extreme weather events. Based on the baseline driving time, Indicates the travel sensitivity coefficient. This indicates real-time precipitation or wind intensity index (unit: mm / hour or standard wind force level). Indicates the circuit breaker threshold for extreme weather events (unit: ) (Keep in line)

[0107] S113. Calculate the equivalent time consumption of real-time dynamic environmental resistance:

[0108] ;

[0109] in, This represents the final output, which is the equivalent time taken to generate real-time dynamic environmental resistance. This indicates the equivalent time spent in real-time traffic congestion. This indicates the equivalent time taken for resistance during extreme weather events.

[0110] The system quantifies students' participation and stability by tracking their total historical attendance within a set period. For example, the system can track the total number of courses students actually attended in the past month or quarter. The instructor's historical performance benchmark rate represents the percentage of classes an instructor actually taught in past schedules, reflecting their reliability and stability. For example, the system can calculate the ratio of the number of classes an instructor actually taught in the past year to the total number of classes scheduled. The instructor's historical average check-in delay quantifies the average delay in check-in before class starts, reflecting their time management and punctuality. For example, the system can record the difference between the instructor's check-in time and the class start time for each class and calculate its average. The venue's current cumulative load duration represents the length of time the venue has been continuously or cumulatively used before the current time. For example, the system can track the total usage time of the venue since the last maintenance or load clearing in real time.

[0111] Among them, the equivalent time of real-time traffic congestion This indicates the additional travel time caused by real-time traffic conditions. Its purpose is to quantify the impact of the external environment on the timely arrival of trainees and instructors. For example, by calling a third-party map service API, the expected arrival time from the trainee / instructor's current location to the training venue under real-time traffic conditions can be obtained. and compared with the baseline driving time Comparison, combined with the penalty coefficient for unintended road blockage Calculations are performed (extracting the ratio of the length of the congested section due to the accident to the total planned route length, assuming the route is clear and there are no unforeseen events). Strictly zeroed out). Equivalent time of extreme weather drag. This indicates the additional travel time caused by extreme weather (such as heavy rain, heavy snow, and strong winds). Its purpose is to quantify the impact of severe weather on travel efficiency and safety. For example, real-time precipitation or wind intensity index can be obtained through a meteorological bureau interface. And combined with travel sensitivity coefficient (As the system runs, the virtual agent, through its preference learning capabilities, collects actual travel delay data from instructors during historical severe weather conditions, and analyzes...) In addition to dynamic calibration and correction, value assignment can also be achieved through expert experience. For example, an expert experience rule base can be introduced to assign travel sensitivity based on the basic commuting patterns registered by coaches or trainees in their personal agent systems. For fully protected commuting modes (such as subways and private cars), the value range is set to 0.10 to 0.25, with the system's preferred default value being 0.15; for semi-protected commuting modes (such as buses and shared shuttles), the baseline travel sensitivity value range is set to 0.35 to 0.60, with the system's preferred default value being 0.45; for unprotected commuting modes (such as walking, cycling, and electric bicycles), the baseline travel sensitivity value range is set to 0.75 to 1.20, with the system's preferred default value being 0.85. (This also includes the threshold for triggering circuit breakers in extreme weather events.) Calculations are performed. Real-time dynamic environmental resistance equivalent time consumption. It is the equivalent time of real-time traffic congestion. Equivalent time to extreme weather drag The sum of these factors reflects the total additional burden on travel time caused by the current external environment.

[0112] Specifically, when extracting multidimensional historical behavioral features, the system no longer simply collects historical data in a general way. Instead, it explicitly extracts from historical logs the total number of students' actual attendances within a set period, the historical performance benchmark rate and average check-in delay time of coaches, and the current cumulative load time of the venue. This specific and quantifiable historical behavioral data provides a solid and refined foundation for students' willingness and preference scores, coaches' time preference scores, and the venue's expected turnover rate. For example, the total number of students' actual attendances directly reflects their level of engagement with the course, the historical performance benchmark rate and average check-in delay time of coaches are directly related to their reliability and time management habits, and the current cumulative load time of the venue directly affects its availability and maintenance needs. In this way, the virtual agent can build its preference model based on more specific and reliable historical data. As a result, in subsequent negotiations, its concession strategies and adjustments to expected utility values ​​will be closer to reality, avoiding negotiation biases caused by data ambiguity. Meanwhile, to address the dynamic uncertainties of the external environment, this invention introduces the concept of real-time dynamic environmental resistance equivalent time and quantifies it precisely. This time is derived from the real-time traffic congestion equivalent time. Equivalent time to extreme weather drag It consists of two parts. Real-time traffic congestion equivalent time By comparing the expected arrival time under the current real-time traffic conditions Compared with the baseline driving time And consider the penalty coefficient for unexpected road blockage. This was used to calculate and quantify the impact of traffic conditions on travel time. The equivalent travel time due to extreme weather drag was also included. Then the benchmark driving time Travel sensitivity coefficient Real-time precipitation or wind intensity index and the intensity of extreme weather events triggering circuit breakers. This is used to calculate and quantify the impact of severe weather on travel efficiency. Adding these two factors together yields the final real-time dynamic environmental drag equivalent time. This refined real-time environmental resistance assessment directly impacts the calculation of the real-time performance prediction rate, enabling the virtual agent to fully consider the impact of the external environment on the likelihood of performance during negotiation. For example, when there is traffic congestion or severe weather, This will increase, leading to a decrease in real-time fulfillment prediction rate. This reduction prompts virtual agents to adjust their utility expectations for that period during negotiations, or to reflect an aversion to environmental risks in their concession strategies.

[0113] The following is a concrete example. Suppose a sports training institution needs to book a badminton lesson for a student. When extracting multidimensional historical behavioral features, the system first retrieves data from historical logs. For example, the student's total actual attendance over the past three months is 10 times; the coach's historical attendance rate is 95%, and the historical average check-in delay time is 5 minutes; the current cumulative load duration of the venue (badminton court A) is 20 hours. This data is accurately extracted and used as the basis for subsequent calculations. When calculating the equivalent time of real-time dynamic environmental resistance, the system obtains external data in real time. For example, it obtains real-time traffic conditions from the coach's home to the training venue through the Gaode Map API, and the expected arrival time under the current real-time traffic conditions. It takes 30 minutes, while the base driving time is... The penalty coefficient for unexpected road blockage is 20 minutes. Set to 0.1. This will determine the equivalent real-time traffic congestion time. =11 minutes. Meanwhile, real-time precipitation was obtained via the Weather Bureau API. The travel sensitivity coefficient is 5 mm / hour. Set to 0.05, the threshold for triggering circuit breakers in extreme weather events. Set to 20 mm / hour. Then the equivalent time for extreme weather drag is... =0.284 minutes. Ultimately, the equivalent time for real-time dynamic environmental resistance is... =11.284 minutes. These specifically quantified multidimensional historical behavioral characteristics and the equivalent time of real-time dynamic environmental resistance will serve as inputs for subsequent quantitative derivation of student willingness and preference scores, coach time preference scores, real-time performance prediction rates, and venue expected turnover rates, thus constituting the multidimensional preference and state parameters of each virtual agent. For example, a student's total historical attendance of 10 times will directly affect the calculation of their willingness and preference score, while the equivalent time of real-time dynamic environmental resistance of 11.284 minutes will directly affect the coach's real-time performance prediction rate. In this way, the system can provide a reliable basis for intelligent negotiation based on accurate data.

[0114] By defining specific data structures and precisely quantifying the equivalent time consumption of multidimensional historical behavioral characteristics and real-time dynamic environmental resistance, this approach solves the problem in intelligent negotiation and appointment methods where the ambiguity or inaccuracy of basic data definitions leads to distortions in virtual agent preferences and state parameters, thus affecting negotiation efficiency and the rationality of the final scheduling. By clearly defining the total number of students' historical attendance, the coach's historical performance benchmark rate and historical average check-in delay time, and the current cumulative load duration of the venue, the virtual agent can build its preference model based on more specific and reliable historical data. This allows its concession strategies and utility expectations to be more closely aligned with reality during negotiation. Furthermore, by introducing real-time traffic congestion equivalent time consumption and extreme weather resistance equivalent time consumption and performing precise mathematical model calculations, the system can dynamically assess the impact of the external environment on the likelihood of performance, thus fully considering these external risks during negotiation.

[0115] Through the aforementioned technical solutions, the construction of multidimensional preferences and state parameters of virtual agents becomes more accurate and dynamic. These precise parameters, serving as the basis for negotiation, can more realistically reflect the actual situation of trainees, coaches, and venues, as well as potential risks in the external environment. This enables multi-round dynamic negotiation to be based on more accurate information, and the final calculated global comprehensive negotiation utility index is more valuable. This ensures that the final generated training schedule not only meets the basic needs of all parties but also maximizes their satisfaction and the likelihood of fulfillment, significantly improving the accuracy, reliability, and adaptability of intelligent negotiation and appointment methods.

[0116] Example 2

[0117] Specifically, the steps for quantitatively deriving multidimensional preferences and state parameters by utilizing multidimensional historical behavioral characteristics and the equivalent time consumption of real-time dynamic environmental resistance are as follows:

[0118] S121. Calculate the score of trainee preference:

[0119] ;

[0120] in, This represents the score indicating the student's willingness and preference. This represents the total number of actual attendances in history (in times). Indicates the threshold number of times (in times) for forming a basic habit;

[0121] S122. Calculate the coach's time preference score:

[0122] ;

[0123] in, This indicates the coach's time preference for scoring. This indicates the available free interval time for the coach (in minutes). Indicates the standard duration of a single course session (in minutes);

[0124] S123. Calculate the real-time fulfillment prediction rate:

[0125] , ;

[0126] in, Indicates the real-time fulfillment prediction rate. Indicates the historical performance benchmark rate. This represents the equivalent time consumed by real-time dynamic environmental resistance. This indicates the time taken to reach the tolerance boundary under extreme environmental conditions (in minutes). This indicates the total number of times a coach has actually fulfilled a contract within a set period (in units of times). This indicates the total number of fixed training sessions a coach has scheduled within the same period (in times).

[0127] S124. Calculate the expected site turnover rate:

[0128] ;

[0129] in, This indicates the expected turnover rate of the venue. This indicates the current cumulative load duration (in hours). This indicates the reference value for the physical maintenance cycle of the site's intrinsic facilities (a preset threshold for the duration of condition maintenance required after the site has been used cumulatively (in hours)).

[0130] Among them, the score of trainees' willingness and preference is calculated. The steps aim to quantify students' enthusiasm and willingness to participate in the course. Their purpose is to provide students with a core indicator to measure their acceptance of the course schedule, as shown in the formula. Number of times compared to the preset basic habit formation threshold The relationship between attendance and coach time preference scores is calculated using an exponential decay model, where higher attendance rates result in higher scores, indicating stronger coach willingness. The steps involved quantify a coach's preference for teaching during specific time slots. Their purpose is to provide the coach's virtual agent with an indicator to assess the attractiveness of different class times. This is achieved through the available free time intervals available to the coach, as shown in the formula. Compared to the standard single course duration The degree of matching between the two was calculated using an exponential decay model. The closer the idle interval was to the standard course length, the higher the score, indicating greater attractiveness of that time slot to the coach. Real-time fulfillment prediction rate was calculated. The steps aim to predict the likelihood of an instructor or student fulfilling their course commitments under current real-time environmental conditions. Its purpose is to provide the virtual agent with a dynamic indicator for assessing external environmental risks. This is achieved, as shown in the formula, based on a historical performance benchmark rate. And combined with the real-time dynamic environmental resistance equivalent time consumption Time consumed by the tolerance boundary of extreme environmental resistance The calculation is performed using an exponential decay model, where the greater the environmental resistance, the lower the performance prediction rate. The expected site turnover rate is then calculated. The steps are used to quantify the availability of the site under the current load and the degree to which the system expects it to be used. Their purpose is to provide the site virtual agent with an indicator to measure the efficiency of site resource utilization and maintenance needs. This is achieved, as shown in the formula, by using the current cumulative load duration... Reference values ​​for physical maintenance cycles of site-specific facilities The relationship between the cumulative load duration and the maintenance cycle is calculated using an exponential decay model. The closer the cumulative load duration is to the maintenance cycle, the lower the turnover rate, indicating that the site needs maintenance or should not be overused.

[0131] Specifically, student preference score It focuses on reflecting learners' learning enthusiasm, quantifying it through their historical attendance data, and providing a basis for learners' representatives to express their acceptance of the course during negotiations. Coach time preference score. This focuses on coaches' inherent preferences for different teaching times, calculating the match between available free intervals and standard course durations. This allows coaching agents to intelligently assess and select the most suitable teaching times. Real-time fulfillment prediction rate. The dynamic influence of the external environment is introduced, and by combining historical performance benchmark rates and the equivalent time consumption of real-time dynamic environmental resistance, a common assessment indicator for performance risk is provided for all virtual agents, making the negotiation results more feasible. Expected site turnover rate. From the perspective of site resource utilization and maintenance, the cumulative load-bearing time is quantified to ensure that site agents can balance the needs of resource utilization and facility maintenance during negotiations. The derivation of these quantitative parameters unifies and standardizes the originally discrete and heterogeneous multidimensional historical behavioral characteristics and the equivalent time of real-time dynamic environmental resistance into comparable and operable values. This provides a solid data foundation and decision-making basis for each virtual agent to implement concession strategies, revise and iteratively update the expected value of utility in subsequent rounds of dynamic negotiations, and ultimately calculate the global comprehensive negotiation utility index.

[0132] The following is a concrete example to illustrate this. Suppose the system needs to schedule a sports training session for a student, a coach, and a venue. First, the system calculates the student's preference score. At that time, the system can retrieve the student's total number of historical attendance records. The limit is 40 times, while the system's preset threshold for forming a basic habit is 40 times. For 50 times, according to the formula The student's virtual agent calculates their willingness and preference score. Secondly, the coach time preference score is calculated. At that time, the system identifies a candidate time period and the available free interval time of coaches before and after that time period. It is 70 minutes, while the standard single class duration is... It is 60 minutes, according to the formula. The virtual agent for the coach calculates its preference score for that period. Again, in calculating the real-time performance prediction rate... At that time, the system first determines the coach's historical performance benchmark rate based on historical data. The value is 0.95. Simultaneously, the real-time dynamic environmental resistance equivalent time for the current candidate time period is obtained and calculated through an external interface. The time is 15 minutes, while the system's preset extreme environmental resistance tolerance boundary time is... It is 30 minutes. According to the formula... The system calculates the coach's real-time performance prediction rate for that period. Finally, it calculates the expected turnover rate of the venue. At that time, the system can query the current cumulative load duration of the site. The reference value for the physical maintenance cycle of the site's intrinsic facilities is 150 hours. It is 200 hours. According to the formula... The virtual agents calculate their expected turnover rate. Through the above calculations, each virtual agent obtains its own quantitative preferences and state parameters, which will serve as the basis for subsequent negotiations.

[0133] Through the above technical solution, this invention can effectively transform raw data such as the total number of students' historical actual attendances, the available idle time of instructors, the historical performance benchmark rate, the equivalent time of real-time dynamic environmental resistance, and the current cumulative load capacity of the venue into student willingness and preference scores, instructor time preference scores, real-time performance prediction rates, and venue expected turnover rates with clear quantitative significance. These quantified multidimensional preference and state parameters provide a unified and comparable decision-making basis for each virtual agent in the system, enabling virtual agents to more accurately assess their own acceptance, preference, and performance risk for different course periods.

[0134] Example 3

[0135] Specifically, the process of constructing the hard-constrained joint passage matrix and filtering out the initial candidate time period combinations is as follows:

[0136] S211, Discretize the target reservation time period into A time slice, for a time slice Determine the characteristic elements of the open state of the venue and the characteristic elements of the coach's teaching boundary state respectively;

[0137] The logical determination model for the characteristic elements of the site's open status is as follows:

[0138] ;

[0139] in, The characteristic element representing the open state of the site (a dimensionless Boolean value). Indicates time slice Available idle time before and after (in minutes). Indicates a reference value for the time required to reset the site's intrinsic facilities (in minutes);

[0140] The logical judgment model for the boundary state feature elements of coaching instruction is as follows:

[0141] ;

[0142] in, The teaching boundary state feature element (a dimensionless Boolean value). This indicates the cumulative continuous teaching time of the coach corresponding to the time slot (meaning the cumulative continuous teaching time of the coach within the current scheduling cycle that is connected to the time slot without interruption, in hours). This indicates the coach's intrinsic physical recovery time reference value (in hours). The limit teaching safety factor is represented by the system extracting the coach's historical longest continuous teaching time in a single day without any teaching accidents or negative evaluations, dividing it by the standard coach intrinsic physical recovery time reference value, and constructing a safety boundary mapping value that is strictly between (0,1].

[0143] S212, will and Perform element-wise multiplication to construct the initial candidate time period mask vector:

[0144] Filter mask vector The time slices with a product of 0, and the consecutive time slices with a product of 1 and a corresponding cumulative duration greater than or equal to the standard single course duration, are used as the initial candidate time slot combination.

[0145] Among them, the characteristic elements of the site's open status It is a boolean value used to indicate the time slice. To determine whether the venue meets the mandatory requirements for opening and availability, the system can query the venue management database to obtain the venue's opening status and scheduled classes for a specific time period, and combine this with the preset minimum buffer time required for venue cleaning or equipment cooling. The determination is made based on the reference value for the reset time of the site's inherent facilities.

[0146] Coaching instruction boundary state feature elements It is also a boolean value used to indicate the time slice. Whether an instructor meets the mandatory requirements for teaching primarily depends on their continuous teaching time limits and whether they have other scheduled classes. For example, the system can calculate an instructor's continuous teaching time before and after a specific time slot based on their historical scheduling records and current schedule, and compare this calculation with preset physiological limits and recovery times. Furthermore, instructors can also set their preferred maximum continuous teaching time through their personal schedule management system, and the system will consider these personalized settings when making its decision.

[0147] After determining the characteristic elements of the site's openness status for each time slot... Boundary state feature elements of coaching instruction After that, the system will and Element-wise multiplication is performed to construct the initial candidate time-segment mask vector. Each element in the mask vector represents the corresponding time slice. The combined availability of the venue and coach is calculated, with a product of 1 indicating that the time slice satisfies both the venue and coach hard constraints, and a product of 0 indicating that they do not. This multiplication operation implements the logical "AND" relationship between the venue and coach hard constraints, enabling the rapid identification of time slices that are unavailable both physically and physiologically.

[0148] Subsequently, the system processed the mask vector. The system performs filtering; specifically, it filters out the mask vector. Time slices with a product of 0 are excluded because they do not meet the strict requirements for the venue or coach. Next, the system identifies all consecutive time slice sequences among the remaining time slices with consecutive products of 1, and further filters out those sequences whose cumulative duration is greater than or equal to the standard single-session duration. These sequences are then used as initial candidate time slot combinations. For example, the system can iterate through the mask vector, identify all consecutive "1" sequences, calculate the total duration of each sequence, and if the total duration meets the course requirements, it is used as a candidate time slot combination.

[0149] Through the above technical solution, this invention can efficiently eliminate a large number of time slots that do not meet the basic conditions in the early stages of course booking requests, based on the objective hard constraints of venue and instructors. This pre-filtering mechanism greatly reduces the search space for subsequent dynamic negotiation, avoiding complex utility evaluations and concession strategy calculations for unreasonable time slots, thereby significantly improving the efficiency and accuracy of the entire intelligent negotiation and booking method. It ensures that the candidate time slot combinations entering the negotiation stage are physically and physiologically feasible, providing a solid foundation for subsequent refined preference matching and utility optimization, so that the final generated timetable not only meets the preferences of all parties, but also conforms to the feasibility of actual operation.

[0150] For example, suppose a user requests to book a 60-minute sports training course, with the target time slot being 14:00 to 18:00 on a certain day. The system can discretize this target time slot into 15-minute time slices, resulting in 16 time slices. For each time slice, the system will reset the time reference value based on the venue's inherent facilities. and the availability of the venue, as well as the reference values ​​for the coach's intrinsic physical recovery time. Safety factor of extreme teaching Based on the coaching schedule, the characteristic elements of the venue's open status were calculated. Boundary state feature elements of coaching instruction For example, if the venue needs to reset its equipment between 14:00 and 14:15, then that time slot... The value is 0; if the coach has already taught continuously from 15:00 to 15:15 and reached the safety limit, then the value of that time slot is 0. The value is 0. Subsequently, the system sets the value of all time slices to 0. and Element-wise multiplication is performed to generate the initial candidate time period mask vector. For example, if Given the sequence [1,1,0,1,0,1,1,0,1,1,1,1,1,1,1,1], the system first eliminates all time slices with a value of 0. Then, from the remaining consecutive time slices with values ​​of 1, it identifies sequences with a cumulative duration greater than or equal to 60 minutes (i.e., four 15-minute time slices). For example, eight consecutive time slices starting at 16:00 (i.e., 16:00-18:00) might form a combination of initial candidate time slots that meet the requirements. In this way, the system can efficiently filter out initial candidate time slots that meet the hard constraints, providing high-quality input for subsequent intelligent negotiation.

[0151] Example 4

[0152] Specifically, the evaluation calculates the initial negotiation utility expectation of each virtual agent for the candidate time slot combination, including:

[0153] S221. Calculate the initial negotiation utility expectation of the student's virtual agent:

[0154] ;

[0155] in, This represents the initial negotiation utility expectation of the student's virtual agent. This represents the score indicating the student's willingness and preference. This indicates the time offset (in minutes) between the candidate time period and the most desired target time period. This indicates the student's maximum time offset tolerance (in minutes).

[0156] S222. Calculate the initial negotiation utility expectation of the coach virtual agent:

[0157] ;

[0158] in, This represents the initial negotiation utility expectation of the coach's virtual agent. This represents the preset weighting coefficient. This indicates the coach's time preference for scoring. This represents the historical average check-in delay time (in minutes) extracted from historical behavioral characteristics. This indicates a reference value for standard course buffer time (in minutes).

[0159] S223, Expected initial negotiation utility of the virtual agent for the computational site:

[0160] ;

[0161] in, This represents the initial negotiation utility expectation of the virtual agent for the venue. This indicates the expected turnover rate of the venue. This indicates the available time interval (in minutes) between the candidate time slot and the existing schedule adjacent to the venue. This indicates a reference value for the time required to reset the site's intrinsic facilities (in minutes).

[0162] Among them, the initial negotiation utility expectation of the student's virtual agent is calculated. This calculation measures a learner's initial acceptance of a particular time slot, and its core lies in combining the learner's intrinsic learning willingness with their tolerance for time deviations. Specifically, it includes a learner's willingness-to-preference integral. This reflects the student's historical attendance and engagement; a higher value indicates a stronger willingness to learn. The time offset between the candidate time slot and the most desired target time slot. This quantifies the gap between the candidate time slot and the student's ideal time. A larger deviation generally indicates lower student satisfaction. The student's most desired target time slot can be determined from their preset preferred time in the system, historical booking habits, or through intelligent analysis of their schedule. (Student's maximum time deviation tolerance) This is a key parameter that defines the learner's resilience to time shifts. It can be a fixed value preset by the system or dynamically adjusted based on the learner's historical behavioral data (e.g., the frequency of past exposure to less-than-ideal time periods). An exponential decay function ensures that the learner's expected utility decreases non-linearly as the time shift increases, thus accurately capturing changes in the learner's time sensitivity.

[0163] Calculate the initial negotiation utility expectation of the virtual agent for the coach. This calculation aims to assess a coach's initial acceptance of a candidate time slot, taking into account both the coach's preference for that time slot and their historical reliability. Preset weighting coefficients are used. Used to balance coach time preference scores (Reflecting the match between this time period and the available free time of the coach) and the historical average check-in delay time extracted from historical behavioral characteristics. (Reflecting the coach's punctuality) contribution to total utility. Coach time preference score. The higher the value, the more advantageous the time period is for the coach. Historical average check-in delay time. The lower the value, the higher the coach's reliability in fulfilling their contractual obligations. (Preset weighting coefficient) This can be a global parameter preset by a system administrator, or it can be personalized based on the coach's qualifications, popularity, or specific course type. Standard course buffer time reference value. This method quantifies the negative impact of delay time on coach utility. It can be an industry-standard value or customized according to different sports or coaching styles, ensuring that the coach's actual performance capabilities and habits are fully considered when calculating utility. The model comprehensively reflects the coach's overall intentions through a combination of weighted summation and exponential decay.

[0164] Initial negotiation utility expectation of virtual agent for computing site This calculation assesses the initial acceptance of a venue for a candidate time slot, primarily considering the venue's expected turnover rate and the vacancy interval between that time slot and the venue's existing schedules. Expected venue turnover rate This reflects the efficiency of venue resource utilization; the higher the turnover rate, the more likely the venue is to accept that time slot. The vacancy time interval between the candidate time slot and existing schedules adjacent to the venue. This measures whether there is sufficient buffer time before and after the specified period for facility reset or preparation. Reference values ​​for site-specific facility reset time. This is the minimum buffer time required for the normal operation of the site. It can be a fixed value set by the site manager based on the type of facility and maintenance needs, or it can be dynamically adjusted according to the intensity of site use or seasonal factors. Through an exponential decay function, when the vacancy interval is insufficient to meet the reset time requirement, the expected utility of the site will be significantly reduced, thereby ensuring the rationality and sustainability of site operation.

[0165] Specifically, for each generated candidate time slot combination, the system will calculate the initial negotiation utility expectation of the student's virtual agent. Initial negotiation utility expectation of the virtual coach agent and the initial negotiation utility expectation of the venue virtual agent. Initial negotiation utility expectation of student virtual agents The calculation incorporates scores of trainees' willingness and preferences. and the time offset between the candidate time slot and the student's most desired target time slot And determined by the student's maximum time offset tolerance. Adjustments were made to ensure both the learners' intrinsic motivation and their flexibility regarding time. The initial negotiation utility expectation of the virtual coaching agent. The calculation is performed using preset weighting coefficients. Balanced coach time preference scoring Compared with the historical average check-in delay time extracted from historical behavioral characteristics The impact of this, and the introduction of standard course buffer time reference values. This comprehensively reflects the coach's satisfaction with the time slot and their reliability in fulfilling their obligations. The initial negotiation utility expectation of the virtual agent at the venue. The calculation incorporates the expected turnover rate of the site. The vacancy time interval between the candidate time slot and the existing class schedule adjacent to the venue. And refer to the reference value for the reset time of the site's inherent facilities. An assessment will be conducted to ensure the rational use of site resources and the needs for facility maintenance.

[0166] The following example illustrates this. Assume the system has generated a candidate time slot combination, such as a Tuesday afternoon from 3:00 PM to 4:00 PM. For the virtual agent representing the student, assume their student preference score is... The value is 0.85 (indicating high student motivation). Their preferred target time slot is Tuesday afternoon from 2:00 PM to 3:00 PM. What is the time offset between this candidate time slot and the preferred target time slot? The maximum time deviation tolerance for trainees is 60 minutes. If the timeframe is set to 90 minutes, then the initial negotiation utility expectation of the student's virtual agent is... This can be calculated as 0.436. For the virtual coach agent, assume their coach time preference score... The value is 0.7 (indicating a generally low match between this time period and the coach's free time), representing the historical average check-in delay time extracted from historical behavioral characteristics. For 10 minutes, if the preset weighting coefficient is... The value is 0.7, which is a reference value for standard course buffer time. If the duration is 20 minutes, then the expected initial negotiation utility of the coach's virtual agent is... This can be calculated as 0.672. For a virtual venue agent, assume its expected venue turnover rate... A value of 0.9 (indicating high site utilization efficiency) represents the time interval between this candidate time slot and any existing scheduled classes adjacent to the site. The time is 10 minutes, which is the reference value for the reset time of the site's inherent facilities. If the timeframe is 15 minutes, then the expected initial negotiation utility of the virtual agent at the venue is... This can be calculated as 0.462. Through the above calculation, the system obtains the initial expected utility values ​​of this specific candidate time slot combination for the students, coaches, and venue, providing a quantitative basis for subsequent negotiations.

[0167] Through the above technical solution, this invention effectively solves the problem of lacking objective and quantitative initial evaluation of candidate time slot combinations by virtual agents of all parties in intelligent negotiation and reservation methods. After generating candidate time slot combinations, by accurately calculating the initial negotiation utility expectations of student virtual agents, coach virtual agents, and venue virtual agents, this invention provides a solid and multi-dimensional evaluation benchmark for the subsequent dynamic negotiation process. This makes negotiation no longer a blind trial, but a quantitative starting point based on the real preferences and constraints of all parties. Specifically, the initial utility expectation of students considers their learning willingness and time flexibility, the initial utility expectation of coaches takes into account their time preference and contract fulfillment reliability, and the initial utility expectation of venues balances their turnover efficiency and facility maintenance needs. This comprehensive initial evaluation can significantly improve the efficiency and success rate of negotiation, avoiding negotiation deadlocks or inefficient cycles caused by information asymmetry or unclear initial expectations. Finally, by providing candidate solutions that better match the initial intentions of all parties, this invention can enable the system to converge more quickly to a scheduling result acceptable to all parties, thereby improving the overall user satisfaction and system operating efficiency of intelligent negotiation and reservation for sports training courses.

[0168] Example 5

[0169] Specifically, the process of calculating the dynamic concession tolerance of each virtual agent and adjusting the expected utility value for iterative updates includes:

[0170] S31, in the During the negotiation rounds, for the virtual coach agent, the coach's dynamic concession tolerance is calculated by combining its cumulative continuous teaching hours with the real-time performance prediction rate, and its iterative utility value is updated accordingly.

[0171] , ;

[0172] in, This indicates the coach's dynamic tolerance for concessions. This indicates the coach's tolerance for basic concessions. This indicates the coach's cumulative continuous teaching time (the total continuous teaching time of the coach within the current scheduling cycle that is consecutive with the candidate time slot without interruption, in hours). This indicates the coach's intrinsic physical recovery time reference value (in hours). Indicates the real-time fulfillment prediction rate. Indicates the historical performance benchmark rate. Indicates the coach virtual agent number Round iteration utility value, This represents the initial negotiation utility expectation of the coach's virtual agent;

[0173] S32. For the site virtual agent, calculate the site's dynamic concession tolerance based on the site's current cumulative load duration, and update its iterative utility value:

[0174] , ;

[0175] in, Indicates the tolerance for dynamic concessions on the field. Indicates the tolerance for site foundation concessions. This indicates the current cumulative load duration. This indicates the reference value for the physical maintenance cycle of the site's intrinsic facilities (in hours). Indicates the current round of negotiation interaction (in seconds). This indicates the maximum number of negotiation rounds (in units of rounds). Indicates the venue virtual agent number Round iteration utility value, This represents the initial negotiation utility expectation of the virtual agent for the venue;

[0176] S33. For virtual agents of students, introduce a willingness-to-preference integral to calculate the student's dynamic concession tolerance and update its iterative utility value:

[0177] , ;

[0178] in, Tolerance should be reduced to accommodate student dynamics. This indicates the student's basic tolerance level for concessions. This represents the score indicating the student's willingness and preference. Indicates the current round of negotiation interaction (in seconds). This indicates the maximum number of negotiation rounds (in units of rounds). Indicates the number of virtual agents for students. Round iteration utility value, This represents the initial negotiation utility expectation of the student's virtual agent.

[0179] Specifically, in the first During the round of negotiations, regarding the virtual coaching agent, their cumulative continuous teaching hours were discussed. With real-time fulfillment prediction rate Combined with calculation of coach dynamic concession tolerance and update its iterative utility value. Among them, the coach's dynamic concession tolerance. This indicates the degree to which the coach's virtual agent is willing to accept a reduction in its utility value in the current negotiation round. The concept quantifies the coach's flexibility in scheduling lessons within a specific context. The coach's cumulative continuous teaching hours are also considered. This refers to the total uninterrupted, scheduled or hypothetical teaching hours of a coach within the current scheduling cycle, consecutive with the current candidate time slot. Its purpose is to reflect the coach's fatigue level and is a key physiological indicator influencing their willingness to compromise. This duration can be obtained through real-time statistics of the coach's scheduling records. For example, by querying the coach's scheduled timetable in the database and calculating the total duration of consecutive classes adjacent to the current candidate time slot, or by accumulating the teaching status data reported in real-time by the coach's application. Coach's basic compromise tolerance. The inherent willingness to make concessions by virtual agents of coaches under ideal or baseline conditions is a preset constant that reflects the average or minimum concession tendency of the coaching group. This baseline tolerance can be set based on the coach's qualifications, contract type, or the system administrator's experience, or by statistically analyzing historical negotiation data to determine the average level of concessions made by coaches without the influence of special external factors.

[0180] For site virtual agents, consider the current cumulative load duration of the site. Calculate the dynamic concession tolerance of the site and update its iterative utility value. Among them, the tolerance for dynamic concessions on the field. This indicates the degree to which the virtual agent of the venue is willing to accept a reduction in its utility value in the current negotiation round. The concept quantifies the flexibility of the venue's usage arrangements in a specific context. Venue-based concession tolerance. The inherent willingness of the virtual agent to make concessions under ideal or baseline conditions is a preset constant that reflects the average or minimum concession tendency of the site management. This basic tolerance can be set according to the type of site, operating strategy, or the experience value of the system administrator; or, by analyzing historical site usage data and negotiation records, the average level of concessions made by the site under normal operating conditions can be obtained.

[0181] For virtual agents of students, a willingness and preference score is introduced. Calculate the student's dynamic concession tolerance and update its iterative utility value. Among them, the student's dynamic concession tolerance. This indicates the degree to which a student's virtual agent is willing to accept a reduction in their utility value in the current negotiation round. The concept quantifies the student's flexibility regarding course scheduling in a specific context. Student Base Concession Tolerance Coefficient This represents the inherent willingness of virtual agents to make concessions under ideal or baseline conditions. It is a preset constant that reflects the average or minimum concession tendency of the student group. This basic tolerance coefficient can be set based on the student's membership level, course type, or the system administrator's experience level. Alternatively, it can be determined by statistically analyzing user survey data and historical appointment records to derive the average concession level of students under the influence of no special external factors. Update and iterate the utility value. , , This refers to adjusting the expected utility value of each virtual agent in each round of negotiation based on the calculated dynamic concession tolerance. Its purpose is to simulate the agent's behavior of adjusting its expected return according to its own state and changes in the external environment during the negotiation process. This update process can be achieved by multiplying the initial expected utility value by the concession tolerance. The greater the tolerance for concessions, the greater the decrease in utility value, indicating a larger concession.

[0182] The following is a concrete example to illustrate this. After the negotiation system is activated, the student's willingness and preference score is first calculated based on the historical behavioral characteristics of the student, coach, and venue, as well as the equivalent time of real-time dynamic environmental resistance. Coach time preference score Real-time fulfillment prediction rate and expected turnover rate of the venue Subsequently, for a course booking request, the system generates a series of candidate time slot combinations and calculates the initial negotiation utility expectation for each virtual agent for these candidate time slot combinations. , , When negotiations entered the first round ( When =1), the system will calculate the dynamic concession tolerance of each virtual agent for each candidate time period combination. For example, for a certain coach virtual agent, assuming its current cumulative continuous teaching time is... Real-time fulfillment prediction rate for 3 hours The historical performance benchmark rate is 0.95. The intrinsic physical recovery time of a coach is 0.90. For 4 hours, the coach's basic concession tolerance The value is 0.1. The system substitutes these parameters into the formula. Calculate the coach's dynamic concession tolerance. Subsequently, using this Initial negotiation utility expectation with the virtual agent of the coach Through formula Calculate the utility value of the coach virtual agent in the first round of iterations. Similarly, for a site virtual agent, assume its current cumulative load duration... Reference value for the physical maintenance cycle of site-specific facilities: 50 hours. The current round of negotiation and interaction lasts for 100 hours. The maximum number of negotiation rounds is 1. The site foundation concession tolerance is 10. The value is 0.05. The system substitutes these parameters into the formula. Calculate the site dynamic concession tolerance and update its iterative utility value. For the virtual agent of the student, assume their willingness and preference integral. The value is 0.8, representing the current round of negotiation. The maximum number of negotiation rounds is 1. The student's basic concession tolerance coefficient is 10. The value is 0.15. The system substitutes these parameters into the formula. Calculate the student's dynamic concession tolerance. and update its iterative utility value. This process is repeated in each round of negotiation until the overall negotiation utility index reaches a preset threshold, at which point the negotiation converges. Through this dynamic adjustment, the expected utility values ​​of each virtual agent gradually converge as the negotiation progresses, thereby enabling the system to find a course schedule acceptable to all parties.

[0183] Through the above technical solution, this invention can effectively solve the problems of rigidity and stagnation that may occur in multi-agent negotiation. During the negotiation process, each virtual agent no longer simply adheres to its initial preferences, but can dynamically adjust its willingness to concede based on its own real-time status (such as the coach's fatigue level, the venue's load) and the progress of the negotiation (such as the negotiation rounds). This mechanism makes the negotiation process more flexible and can simulate the compromise behavior of participants in the real world when facing conflict, thereby significantly improving the efficiency and success rate of negotiation.

[0184] Example 6

[0185] Specifically, the formula for calculating the overall comprehensive negotiation utility index is as follows:

[0186] ;

[0187] in, Indicates the first The overall comprehensive negotiation utility index of this candidate time period combination in the round of negotiation. , , These represent the first, second, and third virtual agents of students, coaches, and venues, respectively. Round iteration utility value.

[0188] Specifically, the overall comprehensive consultation effectiveness index It is a comprehensive quantitative indicator used to measure the performance of a company in the first half of the year. In a round of negotiation, the overall satisfaction or utility level of a particular candidate time slot combination for all participating virtual agents (students, coaches, venues) reflects the degree to which the time slot combination is accepted by all parties under the current negotiation state. It is a key basis for judging whether the negotiation has converged and an important standard for evaluating the merits of different candidate time slot combinations.

[0189] For example, suppose in the first... During the round of negotiation, after a certain candidate time slot combination is adjusted by the dynamic concession strategies of each virtual agent, the student's virtual agent's [number]th [period / time slot] becomes [number]. Round Iteration Utility Value The value is 0.8, the first of the virtual coach agent. Round Iteration Utility Value The value is 0.7, the first of the site virtual agent Round Iteration Utility Value The value is 0.9. At this point, the global comprehensive negotiation utility index for this candidate time period combination is... The calculated value is 0.795. The system compares this result with a preset global utility threshold to determine if the negotiation has converged. If the arithmetic mean is used, the result would be (0.8 + 0.7 + 0.9) / 3 = 0.8, which may not fully reflect the coach's relatively low satisfaction level. The geometric mean, with a value of 0.795, more accurately reflects the overall equilibrium satisfaction, prompting the system to potentially optimize further in subsequent negotiations to improve the coach's satisfaction and achieve a better global solution.

[0190] Through the above technical solution, this invention effectively solves the problem of comprehensively evaluating the utility values ​​of each virtual agent in multi-party negotiation. By using a geometric mean coupling method, it ensures that the global comprehensive negotiation utility index accurately reflects the balance of interests among all parties, avoiding excessive unfairness to any party in the negotiation result. This allows the system to more accurately determine the timing of negotiation convergence and select a final scheduling scheme with high acceptance among all participants, thereby significantly improving the fairness, stability, and efficiency of intelligent negotiation and reservation of sports training course times.

[0191] Example 7

[0192] Specifically, the control model for inversely mapping the final global comprehensive negotiation utility index into the final target notification distribution volume is as follows:

[0193] When the global comprehensive negotiation utility index reaches a preset threshold and the iteration terminates, and this threshold is determined as the final global comprehensive negotiation utility index, the following control calculations are performed:

[0194] ;

[0195] in, This represents the number of final target notifications (in seconds per minute) generated by the system to control the scheduling of underlying channel resources. This represents the system's preset baseline distribution frequency (in times per minute). This represents the final, comprehensive, and negotiated utility indicator.

[0196] Among them, the final overall comprehensive consultation utility index This is a comprehensive quantitative assessment of the quality of the results of multiple rounds of negotiation. It reflects the overall satisfaction or acceptance of the final selected time period by the virtual agents of the trainees, coaches, and venue. A higher value generally indicates that the negotiation results better meet the preferences and constraints of all parties, meaning better negotiation quality. (Final goal notification distribution volume) This is a command generated by the system based on the quality of the negotiation results. It controls the scheduling of underlying communication channel resources and determines the frequency or priority of distributing scheduling commands to corresponding terminals. Its function is to ensure that the negotiation results can be conveyed to trainees, coaches, and venue managers in a timely and effective manner. This distribution volume can be directly mapped to the message queue sending rate, such as how many notifications are sent per minute, or converted into notification priority. Higher distribution volume corresponds to higher priority, ensuring that notifications are processed and delivered first. Basic distribution frequency benchmark value. This is a preset system parameter representing the basic frequency of notification distribution under normal circumstances. It provides a benchmark reference point for the notification distribution volume. A fixed value can be preset based on system design and communication bandwidth, such as 10 times per minute, or it can be dynamically adjusted based on the average notification demand according to historical data, but it still serves as the initial benchmark. Reverse mapping the final global comprehensive negotiation utility index into the final target notification distribution volume is a mechanism that transforms the quality of the negotiation result (utility index) into actual operational instructions (notification distribution volume). Its purpose is to dynamically adjust the urgency of notifications and resource allocation based on the quality of the negotiation result, thereby optimizing the overall system operating efficiency and user experience. This can be achieved by using mathematical models such as exponential functions, linear functions, or piecewise functions, taking the utility index as input and outputting the distribution volume, or by using a lookup table method to preset different distribution volume levels based on the range of the utility index.

[0197] Specifically, the system uses a control model to... Reverse mapping to the final target notification distribution volume Its core idea is that when the quality of the negotiation result (i.e. A higher value indicates higher satisfaction among all parties regarding the time period, and a greater likelihood of fulfillment. In this case, the system can appropriately reduce the notification distribution volume to avoid excessive consumption of system resources. Conversely, if... A lower satisfaction level indicates that the negotiation outcome is not optimal and the satisfaction of all parties is low. In this case, the system needs to invest more resources to distribute scheduling instructions at a higher frequency or priority to ensure timely and accurate information delivery, thereby mitigating the performance risks that may arise from lower satisfaction. This mechanism creates a closed loop for the entire intelligent negotiation and appointment system: from multidimensional preference analysis to dynamic negotiation, and finally to the intelligent distribution of the final result.

[0198] Through the above technical solution, the final global comprehensive negotiation utility index is inversely mapped into the final target notification distribution volume. This solution can intelligently adjust the notification strategy of scheduling instructions based on the quality of the negotiation results. When the utility index of the negotiation results is high, it indicates that all parties are highly satisfied with the time period, and the system can notify at an appropriate frequency to avoid wasting resources. Conversely, when the utility index is low, the system will increase the notification distribution volume, reminding relevant parties more frequently or urgently, thereby effectively reducing the performance risk caused by unsatisfactory negotiation results and ensuring the smooth progress of the training course.

[0199] In summary, this invention, by introducing an adaptive negotiation mechanism, effectively solves the technical problems of traditional sports training course reservation systems in terms of dynamic balancing of multi-party demands, resistance to dynamic environmental resistance, handling of multi-constraint conflicts, and notification distribution control. It provides a more intelligent, efficient, and robust method for negotiating and reserving sports training course times, significantly improving the intelligence level of scheduling and user experience. The construction of multi-dimensional preferences and state parameters of the virtual agent in this invention is more accurate and dynamic. These precise parameters, as the basis for negotiation, can more realistically reflect the actual situation of students, coaches, and venues, as well as potential risks in the external environment. This allows multi-round dynamic negotiation to be based on more accurate information, and the final calculated global comprehensive negotiation utility index is more valuable. This ensures that the final generated training schedule not only meets the basic needs of all parties but also maximizes the satisfaction and fulfillment probability of all parties, significantly improving the accuracy, reliability, and adaptability of the intelligent negotiation and reservation method. This invention can effectively transform raw data such as the total number of students' historical actual attendance, the available free interval of coaches, the historical performance benchmark rate, the equivalent time of real-time dynamic environmental resistance, and the current cumulative load duration of the venue into student willingness and preference scores, coach time preference scores, real-time performance prediction rates, and venue expected turnover rates with clear quantitative significance. These quantified multidimensional preference and state parameters provide a unified and comparable decision basis for each virtual agent in the system, enabling virtual agents to more accurately assess their own acceptance, preference, and performance risks for different course periods. This invention, in the early stages of course booking requests, efficiently eliminates a large number of time slots that do not meet basic conditions based on objective constraints of venue and instructor. This pre-filtering mechanism greatly reduces the search space for subsequent dynamic negotiation, avoiding complex utility evaluations and concession strategy calculations for unreasonable time slots. This significantly improves the efficiency and accuracy of the entire intelligent negotiation and booking method. It ensures that the candidate time slot combinations entering the negotiation stage are physically and physiologically feasible, providing a solid foundation for subsequent refined preference matching and utility optimization. This results in a final timetable that not only satisfies the preferences of all parties but also aligns with practical feasibility. This invention effectively solves the problem in intelligent negotiation and booking methods of lacking objective and quantitative initial evaluation of candidate time slot combinations by virtual agents of all parties. After generating candidate time slot combinations, it provides a solid and multi-dimensional evaluation benchmark for the subsequent dynamic negotiation process by accurately calculating the initial negotiation utility expectations of student, instructor, and venue virtual agents. This makes negotiation no longer a blind trial but a quantitative starting point based on the true preferences and constraints of all parties.This invention considers the learner's initial utility expectation with their learning willingness and time flexibility, the coach's initial utility expectation with their time preference and contract fulfillment reliability, and the venue's initial utility expectation with their turnover efficiency and facility maintenance needs. This comprehensive initial assessment significantly improves the efficiency and success rate of negotiation, avoiding negotiation deadlocks or inefficient cycles caused by information asymmetry or unclear initial expectations. Ultimately, by providing candidate solutions that better align with the initial intentions of all parties, the system can converge more quickly to a scheduling result acceptable to all parties, thereby improving overall user satisfaction and system efficiency in intelligent negotiation and booking of sports training courses. This invention effectively solves the rigidity and stagnation problems that may occur in multi-agent negotiation. During the negotiation process, each virtual agent no longer simply adheres to its initial preferences but can dynamically adjust its willingness to concede based on its real-time status (such as the coach's fatigue level and the venue's load) and negotiation progress (such as the negotiation rounds). This mechanism makes the negotiation process more flexible, simulating the compromise behavior of participants in the real world when facing conflict, thus significantly improving the efficiency and success rate of negotiation. This invention effectively solves the problem of comprehensively evaluating the utility values ​​of various virtual agents in multi-party negotiations. Through a geometric mean coupling method, it ensures that the global comprehensive negotiation utility index accurately reflects the balance of interests among all parties, avoiding excessive unfairness to any party in the negotiation result. This allows the system to more accurately determine the timing of negotiation convergence and select a final scheduling scheme with high acceptance among all participants, thereby significantly improving the fairness, stability, and efficiency of intelligent negotiation and reservation of sports training course times. This invention inversely maps the final global comprehensive negotiation utility index into the final target notification distribution volume. This scheme can intelligently adjust the notification strategy of scheduling instructions based on the quality of the negotiation result. When the utility index of the negotiation result is high, it indicates high satisfaction among all parties for the time period, and the system can notify at an appropriate frequency to avoid resource waste. Conversely, when the utility index is low, the system will increase the notification distribution volume, reminding relevant parties more frequently or urgently, thereby effectively reducing the performance risk caused by unsatisfactory negotiation results and ensuring the smooth progress of training courses.

[0200] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent negotiation and reservation of physical education practical teaching and training course time, characterized in that: This includes creating virtual agents with preferred learning abilities for students, coaches, and venues within the system, and also includes the following steps: S1. Extract the equivalent time consumption of multidimensional historical behavioral features and real-time dynamic environmental resistance, and quantify and derive the student's willingness preference score, coach's time preference score, real-time performance prediction rate and venue expected turnover rate to form the multidimensional preference and state parameters of each virtual agent. S2. In response to the course reservation request, the target time is discretized, and a hard-constrained joint access matrix is ​​constructed by combining the reference value of the site's intrinsic facility reset time and the reference value of the coach's intrinsic physical recovery time. Candidate time slot combinations are filtered and generated, and the initial negotiation utility expectation of each virtual agent for the candidate time slot combinations is initialized. S3. Trigger multiple rounds of dynamic negotiation. Each virtual agent executes a concession strategy based on its own multidimensional preferences and state parameters. Calculate the dynamic concession tolerance of each virtual agent and update the expected utility value of each virtual agent in this round accordingly. S4. Couple the expected value of each virtual agent's iteration utility in the current round in multiple dimensions, calculate the global comprehensive negotiation utility index, and determine the negotiation convergence when the index reaches the preset global utility threshold, and take the current candidate time period combination as the final feasible solution. S5. Generate the final training course schedule based on the feasible solution, and reverse map the final global comprehensive negotiation utility index output when the negotiation converges into the final target notification distribution amount, and distribute the scheduling instructions to the corresponding terminals according to the final target notification distribution amount.

2. The intelligent negotiation and reservation method for sports practice teaching and training course time according to claim 1, characterized in that: In S1, the multidimensional historical behavioral features include the total number of actual attendances of students within a set period extracted from historical logs, the historical performance benchmark rate of coaches and the historical average check-in delay time, and the current cumulative load duration of the venue; the real-time dynamic environmental resistance equivalent time includes the real-time traffic congestion equivalent time and extreme weather resistance equivalent time obtained through external interfaces, and the sum of the two constitutes the real-time dynamic environmental resistance equivalent time.

3. The intelligent negotiation and reservation method for sports practice teaching and training course time according to claim 1, characterized in that: In step S1, the method for calculating the equivalent time consumption of real-time dynamic environmental resistance includes the following steps: S111, Calculate the equivalent time of real-time traffic congestion: ； in, This indicates the equivalent time spent in real-time traffic congestion. This indicates the expected arrival time under the current real-time traffic conditions. Indicates the baseline driving time. This represents the penalty coefficient for unexpected road blockage; S112. Calculate the equivalent time of drag during extreme weather: ； in, This represents the equivalent time spent due to drag during extreme weather events. Based on the baseline driving time, Indicates the travel sensitivity coefficient. This indicates real-time precipitation or wind intensity index. Indicates the circuit breaker threshold for extreme weather events; S113. Calculate the equivalent time consumption of real-time dynamic environmental resistance: ； in, This represents the final output, which is the equivalent time taken to generate real-time dynamic environmental resistance. This indicates the equivalent time spent in real-time traffic congestion. This indicates the equivalent time taken for resistance during extreme weather events.

4. The intelligent negotiation and reservation method for sports practice teaching and training course time according to claim 1, characterized in that: In step S1, the method for quantitatively deriving multidimensional preferences and state parameters by utilizing multidimensional historical behavioral characteristics and the equivalent time consumption of real-time dynamic environmental resistance includes the following steps: S121. Calculate the score of trainee preference: ； in, This represents the score indicating the student's willingness and preference. This represents the total number of actual attendances in history. This indicates the threshold number of times a basic habit has been formed; S122. Calculate the coach's time preference score: ； in, This indicates the coach's time preference for scoring. This indicates the amount of free time available for the coach. Indicates the standard duration of a single course; S123. Calculate the real-time fulfillment prediction rate: ， ； in, Indicates the real-time fulfillment prediction rate. Indicates the historical performance benchmark rate. This represents the equivalent time consumed by real-time dynamic environmental resistance. Indicates the time consumed by the tolerance boundary of extreme environmental resistance. This indicates the total number of times a coach has actually fulfilled their contractual obligations within a set period. This indicates the total number of fixed training sessions a coach has scheduled within the same period. S124. Calculate the expected turnover rate of the site: ; in, This indicates the expected turnover rate of the venue. This indicates the current cumulative load duration. This indicates a reference value for the physical maintenance cycle of the site's intrinsic facilities.

5. The intelligent negotiation and reservation method for sports practice teaching and training course time according to claim 1, characterized in that: In S2, the reference value for the intrinsic facility reset time of the venue is the minimum buffer time required between two uses of the venue; the reference value for the intrinsic physical recovery time of the coach is the minimum recovery time required after the coach has taught continuously.

6. The intelligent negotiation and reservation method for sports practice teaching and training course time according to claim 1, characterized in that: In S2, the method for constructing the hard-constrained joint access matrix and filtering out the initial candidate time period combinations includes the following steps: S211, Discretize the target reservation time period into A time slice, targeting a time slice Determine the characteristic elements of the open state of the venue and the characteristic elements of the coach's teaching boundary state respectively; The logical determination model for the characteristic elements of the site's open status is as follows: ； in, Elements indicating the open status of a site Indicates time slice Available idle time before and after, This indicates a reference value for the time required to reset the site's intrinsic facilities; The logical judgment model for the boundary state feature elements of coaching instruction is as follows: ； in, Represents the feature elements of the teaching boundary state. This indicates the cumulative continuous teaching time of the coach corresponding to the allocated time slot. This indicates the reference value for the coach's intrinsic physical recovery time. Indicates the safety factor for extreme teaching conditions; S212, will and Perform element-wise multiplication to construct the initial candidate time period mask vector: Filter mask vector The time slices with a product of 0, and the consecutive time slices with a product of 1 and a corresponding cumulative duration greater than or equal to the standard single course duration, are used as the initial candidate time slot combination.

7. The intelligent negotiation and reservation method for sports practice teaching and training course time according to claim 1, characterized in that: In step S2, the method for evaluating and calculating the initial negotiation utility expectation of each virtual agent for the candidate time period combination includes the following steps: S221. Calculate the initial negotiation utility expectation of the student's virtual agent: ； in, This represents the initial negotiation utility expectation of the student's virtual agent. This represents the score indicating the student's willingness and preference. This represents the time offset between the candidate time period and the most desired target time period. This indicates the student's maximum tolerance for time offset. S222. Calculate the initial negotiation utility expectation of the coach virtual agent: ； in, This represents the initial negotiation utility expectation of the coach's virtual agent. This represents the preset weighting coefficient. This indicates the coach's time preference for scoring. This represents the historical average check-in delay time extracted from historical behavioral characteristics. This indicates a reference value for the standard course buffer time; S223, Expected initial negotiation utility of the virtual agent for the computational site: ； in, This represents the initial negotiation utility expectation of the virtual agent for the venue. This indicates the expected turnover rate of the venue. This indicates the available time interval between the candidate time slot and the existing schedule adjacent to the venue. This indicates a reference value for the time required to reset the site's intrinsic facilities.

8. The intelligent negotiation and reservation method for sports practice teaching and training course time according to claim 1, characterized in that: In step S3, the process of calculating the dynamic concession tolerance of each virtual agent and correcting the expected value of the iterative update utility includes the following steps: S31, in the During the negotiation rounds, for the virtual coach agent, the coach's dynamic concession tolerance is calculated by combining its cumulative continuous teaching hours with the real-time performance prediction rate, and its iterative utility value is updated accordingly. ， ； in, This indicates the coach's dynamic tolerance for concessions. This indicates the coach's tolerance for basic concessions. This indicates the coach's cumulative continuous teaching hours. This indicates the reference value for the coach's intrinsic physical recovery time. Indicates the real-time fulfillment prediction rate. Indicates the historical performance benchmark rate. Indicates the coach virtual agent number Round iteration utility value, This represents the initial negotiation utility expectation of the coach's virtual agent; S32. For the site virtual agent, calculate the site's dynamic concession tolerance based on the site's current cumulative load duration, and update its iterative utility value: ， ； in, Indicates the tolerance for dynamic concessions on the field. Indicates the tolerance for site foundation concessions. This indicates the current cumulative load duration. This indicates a reference value for the physical maintenance cycle of the site's intrinsic facilities. Indicates the current round of negotiation interaction. This indicates the maximum number of negotiation rounds. Indicates the venue virtual agent number Round iteration utility value, This represents the initial negotiation utility expectation of the virtual agent for the venue; S33. For virtual agents of students, introduce a willingness-to-preference integral to calculate the student's dynamic concession tolerance and update its iterative utility value: ， ； in, Tolerance should be reduced to accommodate student dynamics. This indicates the student's basic tolerance level for concessions. This represents the score indicating the student's willingness and preference. Indicates the current round of negotiation interaction. This indicates the maximum number of negotiation rounds. Indicates the number of virtual agents for students. Round iteration utility value, This represents the initial negotiation utility expectation of the student's virtual agent.

9. The intelligent negotiation and reservation method for sports practice teaching and training course time according to claim 1, characterized in that: In S4, the formula for calculating the global comprehensive negotiation utility index is as follows: ； in, Indicates the first The overall comprehensive negotiation utility index of this candidate time period combination in the round of negotiation. , , These represent the first, second, and third virtual agents of students, coaches, and venues, respectively. Round-by-round utility value.

10. The intelligent negotiation and reservation method for sports practice teaching and training course time according to claim 1, characterized in that: In S5, the control model that inversely maps the final global comprehensive negotiation utility index into the final target notification distribution amount is as follows: When the global comprehensive negotiation utility index reaches a preset threshold and the iteration terminates, and this threshold is determined as the final global comprehensive negotiation utility index, the following control calculations are performed: ； in, This represents the final target notification distribution amount generated by the system to control the scheduling of underlying channel resources. This represents the system's preset baseline distribution frequency value. This represents the final, comprehensive, and negotiated utility indicator.

Images