An experience-based field technician dispatch method and system

Abstract

The present application relates to the technical field of computer, in particular to a kind of experience-based field technician scheduling method and system, method includes: obtaining scheduling input data, including the directed graph of node and task node, travel time, available technician set, skill domain information, proficiency level, skill domain demand and demand proficiency level, income, task time window, work time window, service duration mapping relationship;According to proficiency level and demand proficiency level, compatible relationship is generated and experience-related service duration is calculated;Under the feasibility constraint, scheduling scheme is constructed, task node unique allocation is satisfied, job route is started from node and returns node, service starting time is determined according to travel time and experience-related service duration to satisfy time window;Heuristic insertion generates initial scheduling scheme;Iterative local search generates neighborhood, temperature parameter probability acceptance, disturbance and cooling;The optimal scheduling scheme is output to maximize the income.The present application can solve the problem of inaccurate time consumption estimation and insufficient plan robustness caused by the fact that the experience / proficiency of technicians is not explicitly reflected in the existing field technician scheduling.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to an experience-based method and system for dispatching field technicians. Background Technology

[0002] In on-site service scenarios such as home appliance repair and home services, service providers typically face practical challenges such as dispersed customer distribution, limited number of technicians, time constraints on tasks, and differences in skill types and proficiency required for tasks. They need to develop executable task allocation and travel arrangements within the planning cycle to improve overall service value and resource utilization efficiency. While existing research on on-site technician scheduling has addressed the comprehensive constraints of skill matching and path / time windows, some solutions still rely on the simplistic assumption that different technicians capable of performing the same task have the same service time. This can lead to biased estimates of total task execution time, affecting the feasibility and fairness of the plan, and even causing problems such as timeouts or resource imbalances. Summary of the Invention

[0003] In view of the above technical problems, the present invention provides an experience-based field technician scheduling method and system to solve the problems of inaccurate time estimation and insufficient plan robustness caused by the failure of existing field technician scheduling to explicitly reflect the impact of technician experience / proficiency on service duration. By establishing a correlation between technician proficiency in the skill domain and the proficiency required for the task and determining the experience-related service duration accordingly, a high-quality scheduling scheme is formed under the premise of meeting time constraints and feasibility constraints, so as to improve the accuracy of the task execution phase and the overall benefit level.

[0004] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part from practice of this disclosure.

[0005] According to one aspect of the present invention, an experience-based field technician dispatching method is proposed, the method comprising: Acquire scheduling input data, which includes: a directed graph containing network nodes and multiple task nodes and the travel time between nodes, a set of available technical personnel, skill domain information, the proficiency level of each technical personnel in the skill domain information, the skill domain requirements and required proficiency levels of each task node, the revenue and task time window of each task node, the working time window of the network node, and the service duration mapping relationship used to determine service consumption time. Based on the proficiency level and the required proficiency level, a compatibility relationship between the technician and the task node is generated, and the experience-related service time for each compatibility combination is calculated in conjunction with the service time mapping relationship. A scheduling scheme is constructed under the condition of satisfying the preset feasibility constraints. The feasibility constraints include at least: the unique allocation of the task node, the operation route of the technician starts from the network node and returns to the network node, and the service start time is determined based on the travel time and the service duration related to experience to satisfy the task time window and the work time window. An initial scheduling scheme is generated using heuristic insertion, which is used to select feasible task nodes from the unserved task nodes and insert them into the work routes of technicians compatible with the task nodes to reduce travel time increments, until no feasible insertions are available. An iterative local search is performed on the initial scheduling scheme. The iterative local search includes: generating a neighborhood scheme through local search; accepting a scheme with probability based on a temperature parameter when the neighborhood scheme is inferior to the current scheme; perturbing the current scheme when it is not accepted; and cooling the temperature parameter during the iteration process until the termination condition is met. The optimal scheduling scheme obtained during the output iteration process is used to maximize the total revenue of the served task nodes while satisfying the feasibility constraints.

[0006] Furthermore, the compatibility relationship between the generation technician and the task node includes: A skill proficiency description is established for each technician, the skill proficiency description being used to indicate the proficiency level of the technician in each skill domain; A skill requirement description is established for each of the task nodes, and the skill requirement description is used to indicate the required proficiency level of the task node in each skill domain; The skill proficiency description is matched item by item with the skill requirement description. When the proficiency level of the technician in each skill domain involved in the task node meets the condition of not being lower than the required proficiency level, the technician is determined to be compatible with the task node; otherwise, it is determined to be incompatible. The compatible or incompatible results are written into the compatibility matrix. The compatibility matrix is ​​used to quickly exclude incompatible combinations of technicians and task nodes through a query method in insertion, exchange, replacement and feasibility judgment.

[0007] Furthermore, when calculating the experience-related service duration for each compatible combination, the following are included: A nominal service duration is determined for each task node, and the nominal service duration is decomposed according to the number of skill domains required by the task node to obtain the basic time consumption corresponding to each skill domain. For each required skill domain of the task node, the proficiency level of the technician in that skill domain and the required proficiency level of the task node in that skill domain are read, and the corresponding adjustment coefficient or adjustment rule is output by the service duration mapping relationship to correct the base consumption time, thereby obtaining the sub-service consumption time of that skill domain. The service duration mapping relationship is configured to make the sub-service consumption time of technicians with higher proficiency levels shorter when meeting the same required proficiency level. The time spent on each of the sub-services is aggregated across all skill domains involved in the task node to obtain the time taken for the technician to perform the experience-related service of the task node.

[0008] Furthermore, the feasibility constraints also include: Each accessed task node has a unique predecessor node and a unique successor node in the corresponding technician's work route to maintain the connectivity of the work route. Each technician can form at most one work route that starts from the network node and returns to the network node. The construction of the scheduling scheme includes maintaining auxiliary time information for each of the work routes. The auxiliary time information includes at least the earliest feasible service start time information and the latest allowed service start time information or time margin information along the work route. When performing cross-route movement, insertion, or replacement operations, the auxiliary time information is used to recursively calculate the affected section to determine whether the adjusted work route meets the task time window and the work time window without traversing the entire work route, and the auxiliary time information is updated synchronously when it is determined to be feasible.

[0009] Furthermore, the heuristic insertion for generating the initial scheduling scheme includes: Establish a set of unserved task nodes and initialize empty work routes for each technician; For each of the task nodes in the set of unserved task nodes, enumerate the candidate insertion positions for inserting the task node into the work route of any technician compatible with the task node, and calculate the change in the travel time of the work route before and after insertion for each candidate insertion position as the travel time increment. At the same time, verify the feasibility of the candidate insertion position in terms of time window and return point according to the feasibility constraints. Select the candidate insertion position with the smallest travel time increment from all feasible candidate insertion positions and perform the insertion, and remove the inserted task node from the set of unserved task nodes; The aforementioned enumeration, calculation, and insertion are performed repeatedly until there are no candidate insertion positions in the set of unserved task nodes that satisfy the feasibility constraints.

[0010] Furthermore, the local search includes a route structure optimization process to reduce the travel time of the work route. The route structure optimization process sequentially performs cross-route relocation, cross-route exchange, intra-route relocation, intra-route exchange, and arc set disconnection and reconnection operations on the candidate schemes according to a preset order. Among them, the cross-route relocation operation is used to remove an assigned task node in one work route and insert it into the target position of another work route; the cross-route exchange operation is used to swap the positions of task nodes between different work routes; the intra-route relocation and intra-route exchange operations are used to change the order of task nodes within the same work route; and the arc set disconnection and reconnection operation is used to select the connecting arc set in the work route and change the connection mode between the connecting arc sets to achieve the rearrangement of the intermediate task node sequence. Moreover, each operation adopts the first improvement criterion. When a feasible movement that reduces the total travel time and satisfies the feasibility constraint is found, the candidate scheme is updated and the search continues from the starting operation in the preset order.

[0011] Furthermore, the local search also includes a benefit improvement process for increasing returns, which includes insertion and replacement operations; wherein: The insertion operation is used to insert task nodes from the set of unserved task nodes into candidate positions on the work route, provided that the feasibility constraints are met. An insertion evaluation value is calculated for each candidate insertion scheme. This evaluation value is determined by the ratio of the nonlinearly amplified benefit of the task node to be inserted to the corresponding travel time increment. Furthermore, a disturbance coefficient randomly generated within a preset fluctuation range is introduced into the insertion evaluation value to change the ranking of the evaluation results in different iterations. Candidate insertion schemes are selected based on the insertion evaluation value, and insertion is performed until no feasible insertion is possible. The replacement operation is used to select the visited task node as the task node to be replaced, under the premise of satisfying the feasibility constraints, and to select the replacement task node from the set of unserved task nodes, so that the incremental benefit after replacement is positive and as large as possible. At the same time, the auxiliary time information is used to achieve rapid judgment in the replacement feasibility verification and replacement evaluation value calculation.

[0012] Furthermore, the determination and updating of the temperature parameter includes: during initialization, setting an initial value of the temperature parameter by logarithmic backpropagation based on the benefit level of the initial scheduling scheme, the preset degradation tolerance ratio, and the preset acceptance probability, so that when the candidate scheme experiences a decrease in benefit relative to the initial scheduling scheme due to the degradation tolerance ratio, it is still accepted with the acceptance probability; during the iteration process, the probability of acceptance is calculated using an exponential decay rule, and the acceptance probability decreases as the benefit difference between the candidate scheme and the current scheme increases, and a decision is made on whether to update the current scheme with the candidate scheme based on the comparison result of a random number and the acceptance probability; the disturbance includes selecting a random starting point for each of the job routes and continuously removing several assigned task nodes, and performing re-insertion processing on the unserved task nodes after removal to restore feasible benefits, while increasing, resetting, or limiting the removal scale according to a preset adjustment rule, and gradually reducing the temperature parameter according to a preset cooling rule; the termination condition includes the number of consecutive unimproved times reaching a preset upper limit and / or the number of iteration rounds reaching a preset upper limit.

[0013] According to a second aspect of this disclosure, an experience-based field technician dispatching system is provided, the system comprising: The scheduling input construction module is used to acquire scheduling input data, which includes: a directed graph containing network nodes and multiple task nodes and the travel time between nodes, a set of available technical personnel, skill domain information, the proficiency level of each technical personnel in the skill domain information, the skill domain requirements and required proficiency levels of each task node, the revenue and task time window of each task node, the working time window of the network node, and the service duration mapping relationship used to determine service consumption time. The compatibility and service duration calculation module is used to generate a compatibility relationship between the technician and the task node based on the proficiency level and the required proficiency level, and to calculate the experience-related service duration of each compatibility combination in combination with the service duration mapping relationship. The feasibility constraint modeling module is used to construct a scheduling scheme under the condition of satisfying preset feasibility constraints. The feasibility constraints include at least: the unique allocation of the task node, the operation route of the technician starting from the network node and returning to the network node, and the service start time determined based on the travel time and the experience-related service duration to satisfy the task time window and the work time window. The heuristic initial solution generation module is used to generate an initial scheduling scheme using heuristic insertion. The heuristic insertion is used to select a feasible task node from the unserved task nodes and insert it into the work route of the technician compatible with the task node to reduce the travel time increment until no feasible insertion is available. An iterative local search optimization module is used to perform an iterative local search on the initial scheduling scheme. The iterative local search includes: generating a neighborhood scheme through local search; accepting the neighborhood scheme based on a temperature parameter when it is inferior to the current scheme; perturbing the current scheme when it is not accepted; and cooling the temperature parameter during the iteration process until the termination condition is met. The optimal solution output module is used to output the optimal scheduling scheme obtained during the iteration process. The optimal scheduling scheme is used to maximize the total revenue of the served task nodes while satisfying the feasibility constraints.

[0014] The technical solution disclosed herein has the following beneficial effects: Compared to schemes that treat service duration as a fixed value or independent of the executor, this disclosure introduces the modeling concept of experience-related service duration, enabling technicians with different skill levels to correspond to different service times for the same task. This improves the accuracy of the actual time taken during the execution phase and reduces the risk of timeouts or idle waste caused by estimation bias. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating an experience-based field technician dispatching method as described in the embodiments of this specification. Figure 2 This is a structural block diagram of an experience-based field technician dispatching system as described in the embodiments of this specification. Detailed Implementation

[0016] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this disclosure more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced with one or more of these specific details omitted, or other methods, components, apparatus, steps, etc., can be employed. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0017] Furthermore, the accompanying drawings are merely illustrative of this disclosure. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0018] This invention provides a method for dispatching field technicians based on experience. (See reference...) Figure 1 The diagram shown is a flowchart illustrating an experience-based field technician dispatching method according to an embodiment of the present invention. This method can be applied to electronic devices such as personal computers, servers, controllers, and display control boards. The method can be executed by a device, which can be implemented in software and / or hardware. Specifically, the method may include the following steps S101-S106: In step S101, scheduling input data is obtained. The scheduling input data includes: a directed graph containing network nodes and multiple task nodes and the travel time between nodes, a set of available technicians, skill domain information, the proficiency level of each technician in the skill domain information, the skill domain requirements and required proficiency levels of each task node, the revenue and task time window of each task node, the working time window of the network node, and the service duration mapping relationship used to determine the service consumption time.

[0019] In step S101, the scheduling input data is used to comprehensively characterize the location relationships, personnel capabilities, task attributes, and time attributes in the field service environment, and to provide a unified data foundation for subsequent route construction, time estimation, and target evaluation. The location relationships are represented by a directed graph G=(V,A), where... The set of nodes, where node 0 corresponds to a network node and the remaining nodes correspond to task nodes; the set of arcs. The set of task nodes can be denoted as the traversable edges between nodes. Used to distinguish between tasks and network nodes. Travel time between nodes is expressed as a parameter. The expression describes the movement time from node i to node j. The set of available technicians is denoted as T, representing the range of personnel who can depart from the network node to perform tasks. Skill domain information is denoted as set S, and proficiency level information is denoted as set L. The proficiency of technicians in the skill domain is represented by a binary matrix. Description: If technician t possesses proficiency level l in skill domain s, then It also allows the use of hierarchical relationships to express that higher-level abilities include lower-level abilities. The skill requirements of task nodes are represented by integer vectors. The description indicates the required proficiency level for task j in skill domain s. The rewards and time constraints for each task are given as a set of parameters. Corresponding benefits and task time window The working time window corresponding to the network node This data is used to limit the departure and latest return time range for technical personnel. Together, these data form the static part of the scheduling input, used to define task location, resource availability, skill supply and demand, and time constraint boundaries.

[0020] Service duration mapping is used to reflect the impact of different proficiency combinations on task service time, and a service duration matrix can be used. Storage, in This indicates that the technician has a proficiency level in a certain skill domain. At that time, the required level to complete this skill domain is The service duration or equivalent time consumption corresponding to the task is output. The service duration of the task in different skill domains can be obtained item by item according to this matrix and aggregated to form a service duration schedule related to the executor's ability. At the implementation level, the mapping relationship can also be generated by multiplier rules organized according to demand level and capability level, which is used to split the actual service duration of the task into the skill domain dimension and correct it according to the skill level difference. When splitting, the following can be introduced: This represents the maximum number of skill domains required in the task set. It is used to normalize the task service duration to the basic amount of a single skill domain, and then combine the mapping relationship to generate new service duration inputs for different skill domains.

[0021] In step S102, based on the proficiency level and the required proficiency level, a compatibility relationship between the technician and the task node is generated, and the experience-related service time for each compatible combination is calculated in conjunction with the service time mapping relationship.

[0022] The process of generating compatibility relationships between technical personnel and task nodes includes: establishing skill proficiency descriptions for each technical personnel, whereby the skill proficiency descriptions indicate the proficiency level of the technical personnel in each skill domain; establishing skill requirement descriptions for each task node, whereby the skill requirement descriptions indicate the required proficiency level of the task node in each skill domain; matching the skill proficiency descriptions with the skill requirement descriptions item by item; determining that the technical personnel are compatible with the task node when the proficiency level of the technical personnel in each skill domain involved in the task node meets the condition of not being lower than the required proficiency level; otherwise, determining that they are incompatible; and writing the compatible or incompatible results into a compatibility matrix, whereby the compatibility matrix is ​​used to quickly exclude incompatible combinations of technical personnel and task nodes through a query method during insertion, exchange, replacement, and feasibility judgment.

[0023] And, when calculating the experience-related service duration for each compatible combination, the process includes: determining a nominal service duration for each task node, and decomposing the nominal service duration according to the number of skill domains required by the task node to obtain the base consumption time corresponding to each skill domain; for each required skill domain of the task node, reading the technician's proficiency level in that skill domain and the required proficiency level of the task node in that skill domain, and outputting the corresponding adjustment coefficient or adjustment rule from the service duration mapping relationship to correct the base consumption time, thereby obtaining the sub-item service consumption time for that skill domain, wherein the service duration mapping relationship is configured to make the sub-item service consumption time shorter for technicians with higher proficiency levels when meeting the same required proficiency level; and aggregating the sub-item service consumption time across all skill domains involved in the task node to obtain the experience-related service duration for the technician to execute the task node.

[0024] As an explanation, in step S102, the compatibility relationship between technicians and task nodes can be characterized by a skill domain set and a proficiency level set. Let the skill domain set be S, and the proficiency level set be L. For each technician... Using a binary matrix Describe their skill proficiency, among which Indicates that technician t is in the skill domain Reach the proficiency level This description typically satisfies the monotonic coverage property: when a technician reaches a certain level, those below that level are also marked as 1, i.e., if... Then for any There are also This is used to express the relationship where higher-level capabilities contain lower-level capabilities. At the same time, for each task node... To create a skill requirement description, a requirement vector can be used. This indicates the minimum proficiency level required for the task in skill domain s. When matching the skill proficiency descriptions of technicians against the skill requirement descriptions of the tasks, the required level is read for each skill domain s involved in the task. And check whether the technicians meet the requirements. When all skill domains involved in the task satisfy this condition, that is, for all relevant s, When a technician t is deemed compatible with a task node j, incompatibility is determined if any skill domain does not meet the requirement. To facilitate rapid exclusion of incompatible combinations during subsequent insertion, exchange, replacement, and feasibility assessments using a query method, the compatibility determination results are written into a binary parameter matrix. ,in This indicates that technician t is able to execute task node j. This indicates that the operation cannot be performed; the matrix is ​​essentially a matrix... with vector The matching results are compressed and rewritten for compatibility checks and filtering with constant time complexity.

[0025] The calculation of experience-related service time is based on the nominal service time of the task, and the impact of skill proficiency on time consumption is introduced through a mapping relationship. The service time of task node j depends on the proficiency level of the technician performing the task in each skill domain; to characterize the impact of different proficiency level combinations on the time consumption of a single skill domain, a service time matrix is ​​introduced. ,in This indicates that the technician has a proficiency level in a certain skill domain. At that time, the required level to complete this skill domain is Output the duration of each service item corresponding to the task, or an equivalent time consumption coefficient. Experience-related service duration. Skills can be accumulated item by item according to the skill domains required for the task: For each skill domain s involved in task node j, read the task requirement level. Then read the technician's actual proficiency level in that skill domain. (can be satisfied) And the values ​​are determined in accordance with the mapping of the level set), and are obtained through the mapping matrix. Obtain the service duration for each item within the skill domain, and then sum them over all skill domains involved in the task. This setting ensures that the same level of demand is met. Under the premise that higher skill levels correspond to shorter service durations, i.e., higher efficiency. Producing smaller Or even shorter output time, thus reflecting the experience-dependent service duration mechanism.

[0026] In the implementation of decomposing and mapping nominal service duration to skill domains, the number of skill domains with the highest demand can be used as a unified decomposition metric. Let... To determine the maximum number of skill domains required across all tasks, first determine the actual nominal service duration for each task (related to the inherent workload of the task node), then calculate the nominal service duration based on... Decomposing the data yields the base time for a single skill domain: Base Time When a task actually involves multiple skill domains, each skill domain is... As the base time input, combined with the technician's proficiency in that skill domain and the required proficiency level for the task in that skill domain, an adjustment coefficient is output through a mapping relationship to correct the base time, thus obtaining the service time for each item in that skill domain. For example, in the case of a proficiency level of four, a set of data can be used according to the required level. With a level Multiplier table of the organization Scaling is applied: when the skill domain requirement for a task is R1 and the technician's skill level is A3, the multiplier is 0.8; when the requirement is R2 and the skill level is A3, the multiplier is 1.0. Therefore, if the task contains two skill domains with corresponding requirements of R1 and R2, and the technician's skill level is A3 in both domains, and... If the time is 5 minutes, then the experience-related service time for this task to the technician can be summed up to obtain 0.8*5 + 1.0*5 = 9 minutes. The same logic applies to Level 6 proficiency; the time is obtained by adjusting the base time for each skill domain using the corresponding multiplier table and then summing the results. This leads to the formation of experience-related service durations for each compatible combination (t,j).

[0027] In step S103, a scheduling scheme is constructed under the condition of satisfying preset feasibility constraints. The feasibility constraints include at least: the unique allocation of the task node, the operation route of the technician starts from the network node and returns to the network node, and the service start time is determined based on the travel time and the experience-related service duration to satisfy the task time window and the work time window.

[0028] The feasibility constraints further include: each accessed task node has a unique predecessor node and a unique successor node in the corresponding technician's work route to maintain the connectivity of the work route; each technician forms at most one work route starting from and returning to the network node; the construction of the scheduling scheme includes maintaining auxiliary time information for each work route, the auxiliary time information including at least the earliest feasible service start time information and the latest allowed service start time information or time margin information along the work route; when performing cross-route movement, insertion, or replacement operations, the auxiliary time information is used to recursively calculate the affected section to determine whether the adjusted work route meets the task time window and the work time window without traversing the entire work route, and the auxiliary time information is updated synchronously when it is determined to be feasible.

[0029] As an explanation, step S102 is used to organize the representation and determination of the scheduling scheme within the feasibility constraint framework, ensuring that task allocation, travel routes, and time arrangements remain consistent under the same set of constraints. Based on a directed graph, nodes include network point nodes and task nodes; the set of technical personnel is T, the set of task nodes is V, and network point nodes are denoted as 0. Binary variables are used. Represents task node Should it be assigned to technical personnel? Using binary variables Indicate whether technician t travels from node i to node j using a continuous variable. This indicates the start time of service provided by technician t at node i; the travel time between nodes is... The service duration related to the experience of technicians at node i is The task time window is The upper limit of the branch's working hours window is The compatibility parameters are Unique task allocation is achieved by imposing an at most one allocation constraint on each task node: ; Route connectivity is constrained by the pairing relationship between in-degree and out-degree, ensuring that each visited node has only one predecessor and one successor, and binding whether a node is visited to whether a corresponding arc connection exists: ; ; Each technician can form at most one work route starting from and returning to a network node, which can be achieved by limiting the number of arcs originating from the network node: ; The service start time is determined based on travel time and experience-related service duration, selecting the arc. At that time, the start time of node j must not be earlier than the start time of node i plus the service time at i and the travel time from i to j; to make this logic compatible with binary arc selection, a large constant linearization is adopted: , ; Where M is a sufficiently large constant, used in The timing constraint is lifted when the task is assigned. The task time window uses upper and lower bound constraints to serve the start time; when a task is assigned, the start time must fall within the allowed range. , ; The branch operation time window can be reflected through a return feasibility constraint, requiring that the service can be completed before the branch closes; to be consistent with the assignment variable, it is written as: , ; Compatibility is used to exclude skill-incompatible assignment relationships, preventing incompatible combinations from entering the feasible region: ; Based on this, to support rapid feasibility assessment of cross-route movements, insertions, and replacements, auxiliary time information is maintained for each work route: the earliest feasible service start time sequence obtained by forward recursion along the route, and the latest allowed service start time sequence or equivalent time margin information obtained by backward recursion along the route; when an operation only changes a local segment in the route, the nodes at both ends of the segment are used as boundaries, utilizing... and Perform forward recursion on the earliest start time of nodes within the segment, and backward recursion on the latest allowed start time. If any node satisfies... If a conflict occurs, or if the feasibility constraint of returning to the network node is violated, it is judged as infeasible; if it is judged as feasible, only the auxiliary time information of the affected section and its subsequent parts is updated, thereby avoiding a full recalculation of the entire route, improving the efficiency of feasibility verification and maintaining the consistency of the scheduling scheme under the task time window and the working time window.

[0030] In step S104, an initial scheduling scheme is generated using heuristic insertion. The heuristic insertion is used to select a feasible task node from the unserved task nodes and insert it into the work route of a technician compatible with the task node to reduce the travel time increment until no feasible insertion is available.

[0031] The heuristic insertion generation initial scheduling scheme includes: establishing a set of unserved task nodes and initializing empty work routes for each technician; for each task node in the set of unserved task nodes, enumerating candidate insertion positions for inserting the task node into any technician's work route compatible with the task node, and calculating the change in travel time before and after insertion for each candidate insertion position as the travel time increment, while verifying the feasibility of the candidate insertion position in terms of time window and return point according to the feasibility constraints; selecting the candidate insertion position with the smallest travel time increment from all feasible candidate insertion positions and performing the insertion, and removing the inserted task node from the set of unserved task nodes; repeating the above enumeration, calculation and insertion until there are no candidate insertion positions in the set of unserved task nodes that satisfy the feasibility constraints.

[0032] As an explanation, in step S103, the initial scheduling scheme is expressed with the technician's work route as the core, and together with a set of auxiliary information, it forms the representation of the solution. This auxiliary information is used to support the efficient verification of the feasibility of the time window, so that several classic operations can be processed with constant time complexity in the verification stage, and this representation system can be reused in subsequent local improvement processes to reduce the overhead caused by repeatedly traversing the route. The generation of the initial scheduling scheme adopts a step-by-step insertion mechanism: all task nodes are initialized as a set of unserved task nodes, and the work route of each technician is initialized as an empty route. The empty route logically satisfies the route framework of starting from the network node and returning to the network node. For candidate task nodes in the set of unserved task nodes, a fast screening is performed based on compatibility relations, and only combinations that can be executed by the corresponding technicians are inserted for evaluation; the compatibility relations can be represented by binary parameters to indicate whether the technician has the ability to meet the task skill requirements, thereby directly excluding unexecutable combinations in the candidate generation stage. When evaluating the insertion of a candidate task node, for each compatible technician, candidate insertion positions between the task node and its adjacent nodes on the current work route are enumerated. For each candidate insertion position, the change in travel time before and after insertion is calculated as the travel time increment, with the travel time cumulatively measured based on the travel time parameters between nodes. Simultaneously, feasibility checks are performed on the candidate insertion positions, covering time window satisfaction and the feasibility of returning from the network node. Time window checks can be performed using auxiliary time information in the solution representation, allowing for a determination of whether the service start time after insertion still falls within the task time window and the network node's working time window simply by recursively updating the affected local segments. When multiple feasible insertion positions exist for the same candidate task node, the insertion method with the best travel time increment is retained. When multiple candidate task nodes have feasible insertion positions, the candidate insertion with the smallest travel time increment is selected from all feasible candidate insertions, thus prioritizing the suppression of the increasing trend of route travel time during the initial solution construction phase. After insertion, the inserted task node is removed from the set of unserved task nodes, and the corresponding technician's work route and its auxiliary time information are updated synchronously according to the insertion result to ensure that the next round of enumeration and verification is based on the latest route status. This insertion process continues to iterate until there are no task nodes in the set of unserved task nodes that can pass the feasibility verification in any technician's work route, at which point the initial scheduling scheme is generated.

[0033] In step S105, an iterative local search is performed on the initial scheduling scheme. The iterative local search includes: generating a neighborhood scheme through local search; accepting the scheme probabilistically based on a temperature parameter when the neighborhood scheme is inferior to the current scheme; perturbing the current scheme when it is not accepted; and cooling the temperature parameter during the iteration process until the termination condition is met.

[0034] The local search includes a route structure optimization process to reduce the travel time of the work route. This process sequentially performs cross-route relocation, cross-route swap, intra-route relocation, intra-route swap, and arc set disconnection and reconnection operations on the candidate solutions according to a preset order. Specifically, the cross-route relocation operation removes an assigned task node from one work route and inserts it into the target position of another work route; the cross-route swap operation exchanges the positions of task nodes between different work routes; the intra-route relocation and swap operations change the order of task nodes within the same work route; and the arc set disconnection and reconnection operation selects a set of connecting arcs in the work route and changes the connection method between these arcs to rearrange the intermediate task node sequence. Furthermore, each operation adopts a first-time improvement criterion: when a feasible movement that reduces the total travel time and satisfies the feasibility constraints is found, the candidate solution is updated, and the search continues from the preset order of initial operations.

[0035] Furthermore, the local search also includes a benefit improvement process for increasing benefits, which includes an insertion operation and a replacement operation; wherein: the insertion operation is used to insert task nodes from the set of unserved task nodes into candidate positions of the work route, provided that the feasibility constraints are met, and to calculate an insertion evaluation value for the candidate insertion schemes. The insertion evaluation value is determined by the ratio of the benefit of the task node to be inserted after nonlinear amplification processing to the corresponding travel time increment, and the insertion evaluation value introduces a disturbance coefficient randomly generated within a preset fluctuation range to change the ranking of the evaluation results in different iterations; a candidate insertion scheme is selected based on the insertion evaluation value and the insertion is performed until no feasible insertion is possible; the replacement operation is used to select visited task nodes as replaced task nodes, provided that the feasibility constraints are met, and to select replacement task nodes from the set of unserved task nodes, so that the benefit increment after replacement is positive and as large as possible, while using the auxiliary time information to achieve rapid judgment in the replacement feasibility verification and replacement evaluation value calculation.

[0036] The determination and updating of the temperature parameter includes: during initialization, setting an initial value of the temperature parameter by logarithmic backpropagation based on the benefit level of the initial scheduling scheme, the preset degradation tolerance ratio, and the preset acceptance probability, so that when the benefit of a candidate scheme decreases relative to the initial scheduling scheme by the degradation tolerance ratio, it is still accepted with the acceptance probability; during the iteration process, the probability of acceptance is calculated by an exponential decay rule, and the acceptance probability decreases as the benefit difference between the candidate scheme and the current scheme increases, and a decision is made on whether to update the current scheme with the candidate scheme based on the comparison result of a random number and the acceptance probability; the disturbance includes selecting a random starting point for each of the job routes and continuously removing a number of assigned task nodes, and performing re-insertion processing on the unserved task nodes after removal to restore feasible benefits, while increasing, resetting, or limiting the removal scale according to a preset adjustment rule, and gradually reducing the temperature parameter according to a preset cooling rule; the termination condition includes the number of consecutive unimproved times reaching a preset upper limit and / or the number of iteration rounds reaching a preset upper limit.

[0037] As an explanation, for the initial scheduling scheme, a current scheme S, a neighborhood scheme Sw, and a historical best scheme S* are constructed and maintained. Through iterative local search, Sw is repeatedly subjected to neighborhood shifts, and S and S* are updated accordingly. Simultaneously, a temperature parameter T is introduced as the core control variable for the probability acceptance criterion and cooling mechanism, enabling the search to overcome local optima near the target benefit. This iterative process is centered on a single solution: in each round, a local search is performed on Sw to obtain an updated Sw; if the updated Sw makes the target value obj(Sw) higher than obj(S), then S is replaced by Sw, and S* is updated synchronously when obj(Sw) is better than obj(S*); when obj(Sw) is lower than obj(S), Sw is not directly discarded, but is probabilistically accepted according to the simulated annealing acceptance rule, i.e., Sw is accepted with probability. Accepted, and by generating a random number r and comparing The result determines whether to update S with Sw.

[0038] The temperature parameter T is initialized using a logarithmic backwards method to an initial value that ensures the degradation scheme still has a preset acceptability probability: T is set to allow the relative initial level to appear 1+ The degraded scheme is still based on probability. The accepted numerical value is calculated in the following form: ,in To control the tolerance ratio for deterioration, This is the initial acceptance probability control variable. During the iteration process, T is gradually reduced according to the cooling rule, using multiplicative cooling. ,in The temperature decay coefficient is used to tighten the overall acceptance probability of the above exponent as the iteration progresses. The termination condition uses the consecutive unimproved count ηmax as the key threshold: the loop stops when the consecutive unimproved count in the outer loop reaches ηmax; the inner loop limits the number of neighborhood generation times at a single temperature level to ηloop, thus forming a two-layer iterative structure and a stopping criterion.

[0039] The process of generating neighborhood solutions through local search comprises two closely linked parts: a route structure optimization process to reduce the travel time of work routes, and a benefit improvement process to enhance profitability. The route structure optimization part employs five classic neighborhood moves executed in a fixed order: cross-route relocation, cross-route exchange, intra-route relocation, intra-route exchange, and arc set disconnection and reconnection. Cross-route relocation corresponds to `relocate-inter`, which removes a task node from one work route and inserts it into the target position of another work route; cross-route exchange corresponds to `exchange-inter`, which swaps the positions of two task nodes between two work routes; intra-route relocation corresponds to `relocate-intra`, which inserts a task node into a new position within the same work route; intra-route exchange corresponds to `exchange-intra`, which swaps the access order of two task nodes within the same work route; and arc set disconnection and reconnection corresponds to `2-opt`, which removes and replaces two connecting arcs in a route and rearranges the sequence of intermediate task nodes between the two arcs. All five types of movements adopt the first-improvement criterion: feasible combinations are traversed, and once a feasible movement that can reduce the total travel time is found, the candidate scheme is updated with the neighborhood result and the neighborhood index is reset. The search starts again from the starting movement in the preset order to ensure that the improvement is absorbed in time and continuously strengthened. In order to ensure that the feasibility of the time window can be quickly determined in cross-route movement and benefit improvement operations, the candidate scheme maintains auxiliary time information in addition to recording the route sequence, which is used to realize the O(1) time window feasibility check under several operators; while in the intra-route movement, the traditional non-constant complexity check method is used to complete the feasibility verification. At the same time, for the matching constraints between the task and the technician, the compatibility judgment in constant time can be realized by pre-stored binary compatibility matrix, thereby avoiding repeated high-overhead matching calculations during neighborhood evaluation.

[0040] The revenue improvement process consists of insertion and replacement operations, opportunistically absorbing unserved task nodes and further increasing total revenue. The insertion operation iteratively inserts unvisited task nodes into the existing work route, selecting insertion targets based on specific criteria: the evaluation criterion is the ratio of the square of the revenue of the task node to be inserted to the change in travel time caused by the insertion; that is, selecting candidate task and location combinations that maximize the "square revenue / increment in travel time" ratio, balancing revenue increase and time cost, and stopping insertion when no feasible insertion can be found. To avoid the evaluation ranking becoming fixed during iteration and getting trapped in local optima, the insertion and replacement metrics can also be multiplied by a random perturbation factor u, where... This allows for subtle changes in the priority of candidate actions across different iterations, thereby increasing the probability of escaping local structure traps. The replacement operation selects the node to be replaced from visited task nodes and the replacement node from unvisited task nodes, ensuring that the total incremental benefit after replacement is positive and as large as possible. The aforementioned auxiliary time information and time window fast check mechanism are reused in the replacement feasibility verification to maintain low decision overhead when evaluating a large number of candidate replacement pairs.

[0041] When a neighboring solution is not accepted in the probabilistic acceptance decision, to push the search away from the current structure, the current solution needs to be perturbed and a new Sw needs to be generated to continue the iteration. The perturbation process uses shake(is) to perform continuous removal on each job route: with is as the removal scale, a starting position is randomly determined in each route and is consecutively visited task nodes are removed; if removal cannot continue at the current starting point, the position wraps back to the starting point of the route to complete the continuous removal process, thereby breaking the local fragment structure while maintaining the overall framework of the route. The unserved task nodes after removal are reinserted through the insertion mechanism to restore the feasible benefit level and re-form the feasible solution structure. The perturbation strength is adaptively adjusted with iteration: adjustValue() gradually increases is according to rules, and resets it to 1 when is exceeds the number of visited tasks of the longest route. It also allows is to be restored to the default value when improvements occur to maintain the balance between perturbation and reinforcement.

[0042] In step S106, the optimal scheduling scheme obtained during the iteration process is output. The optimal scheduling scheme is used to maximize the total revenue of the served task nodes while satisfying the feasibility constraints.

[0043] Step S106 corresponds to the recording and output of solutions after the iterative search. The output object is the scheduling scheme with the maximum objective function value during the iteration process. The quality of the scheduling scheme is determined by maximizing the total revenue of the tasks served. The objective function can be written as: ; in, This represents the reward of task node i. Indicates whether task node i is executed by technician t.

[0044] During the iteration process, a historical optimal solution cache S and its corresponding objective function value are maintained. When the candidate solution generated in the current round is feasible and its objective function value is higher than the recorded optimal value, the candidate solution is used to cover S. Feasibility is based on the premise that the task is assigned at most once, the route starts from the network node and returns within the working time window, and the service starts at the time that the task time window is met. The relevant constraints are verified before the solution is updated to ensure that the solution entering S* can be directly used for execution.

[0045] The output will fully expand the information in S* into an executable result: For each technician, a visit sequence starting from a network node and returning to that node will be provided, where each task node in the sequence satisfies a unique predecessor and a unique successor connectivity structure; the service start time of each task node will also be given. Service duration related to corresponding experience This ensures that adjacent nodes satisfy the time continuity requirement: ; It also satisfies the constraints of the task time window and the latest time to return to the network node, thereby ensuring that the output plan can be implemented during the execution phase.

[0046] When multiple options have the same maximum total benefit, the option that first reaches the maximum benefit can be saved as the output option, or the option with the shorter travel time can be selected as the output option without changing the benefit target, in order to reduce travel costs and buffer time.

[0047] Based on the same line of thought, such as Figure 2 The diagram shown is a structural block diagram of an experience-based field technician dispatching system provided in an embodiment of the present invention. The system includes: The scheduling input construction module 201 is used to acquire scheduling input data, which includes: a directed graph containing network nodes and multiple task nodes and the travel time between nodes, a set of available technical personnel, skill domain information, the proficiency level of each technical personnel in the skill domain information, the skill domain requirements and required proficiency levels of each task node, the revenue and task time window of each task node, the working time window of the network node, and the service duration mapping relationship used to determine the service consumption time. The compatibility and service duration calculation module 202 is used to generate a compatibility relationship between the technician and the task node based on the proficiency level and the required proficiency level, and to calculate the experience-related service duration of each compatibility combination in combination with the service duration mapping relationship. The feasibility constraint modeling module 203 is used to construct a scheduling scheme under the condition of satisfying the preset feasibility constraints. The feasibility constraints include at least: the unique allocation of the task node, the operation route of the technician starts from the network node and returns to the network node, and the service start time is determined based on the travel time and the experience-related service duration to satisfy the task time window and the work time window. The heuristic initial solution generation module 204 is used to generate an initial scheduling scheme using heuristic insertion. The heuristic insertion is used to select a feasible task node from the unserved task nodes and insert it into the work route of the technician compatible with the task node to reduce the travel time increment until no feasible insertion is available. The iterative local search optimization module 205 is used to perform an iterative local search on the initial scheduling scheme. The iterative local search includes: generating a neighborhood scheme through local search; accepting the neighborhood scheme based on a temperature parameter when it is inferior to the current scheme; perturbing the current scheme when it is not accepted; and cooling the temperature parameter during the iteration process until the termination condition is met. The optimal solution output module 206 is used to output the optimal scheduling scheme obtained during the iteration process. The optimal scheduling scheme is used to maximize the total revenue of the served task nodes while satisfying the feasibility constraints.

[0048] Compared to solutions that treat service duration as a fixed value or independent of the executor, this system introduces the modeling concept of experience-related service duration, enabling technicians with different skill levels to have different service times for the same task. This improves the accuracy of the actual time taken during the execution phase and reduces the risk of timeouts or idle waste caused by estimation bias.

[0049] The specific details of the above system have been described in detail in the method section of the implementation plan. For any undisclosed details, please refer to the implementation plan of the method section, and therefore will not be repeated here.

[0050] This system introduces a state-of-charge dependent quadratic hysteresis resistance model and least squares parameter estimation to separate the hysteresis effect from the effective resistance, obtain a matched hysteresis-free internal resistance, improve estimation accuracy, and eliminate the need for high computational resources.

[0051] The accompanying drawings are merely illustrative of the processes included in the methods according to exemplary embodiments of this disclosure and are not intended to be limiting. It is readily understood that the processes shown in the drawings do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0052] It should be noted that although several modules or units of the system have been mentioned in the detailed description above, this division is not mandatory. In fact, according to exemplary embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0053] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0054] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. An experience-based field technician scheduling method, characterized in that, The method includes: Acquire scheduling input data, which includes: a directed graph containing network nodes and multiple task nodes and the travel time between nodes, a set of available technical personnel, skill domain information, the proficiency level of each technical personnel in the skill domain information, the skill domain requirements and required proficiency levels of each task node, the revenue and task time window of each task node, the working time window of the network node, and the service duration mapping relationship used to determine service consumption time. Based on the proficiency level and the required proficiency level, a compatibility relationship between the technician and the task node is generated, and the experience-related service time for each compatibility combination is calculated in conjunction with the service time mapping relationship. A scheduling scheme is constructed under the condition of satisfying the preset feasibility constraints. The feasibility constraints include at least: the unique allocation of the task node, the operation route of the technician starts from the network node and returns to the network node, and the service start time is determined based on the travel time and the service duration related to experience to satisfy the task time window and the work time window. An initial scheduling scheme is generated using heuristic insertion, which is used to select feasible task nodes from the unserved task nodes and insert them into the work routes of technicians compatible with the task nodes to reduce travel time increments, until no feasible insertions are available. An iterative local search is performed on the initial scheduling scheme. The iterative local search includes: generating a neighborhood scheme through local search; accepting a scheme with probability based on a temperature parameter when the neighborhood scheme is inferior to the current scheme; perturbing the current scheme when it is not accepted; and cooling the temperature parameter during the iteration process until the termination condition is met. The optimal scheduling scheme obtained during the output iteration process is used to maximize the total revenue of the served task nodes while satisfying the feasibility constraints.

2. The experience-based field technician dispatching method according to claim 1, characterized in that, The compatibility relationship between the technical personnel and the task node includes: A skill proficiency description is established for each technician, the skill proficiency description being used to indicate the proficiency level of the technician in each skill domain; A skill requirement description is established for each of the task nodes, and the skill requirement description is used to indicate the required proficiency level of the task node in each skill domain; The skill proficiency description is matched item by item with the skill requirement description. When the proficiency level of the technician in each skill domain involved in the task node meets the condition of not being lower than the required proficiency level, the technician is determined to be compatible with the task node; otherwise, it is determined to be incompatible. The compatible or incompatible results are written into the compatibility matrix. The compatibility matrix is ​​used to quickly exclude incompatible combinations of technicians and task nodes through a query method in insertion, exchange, replacement and feasibility judgment.

3. The experience-based field technician dispatching method according to claim 1, characterized in that, When calculating the experience-related service duration for each compatible combination, the following is included: A nominal service duration is determined for each task node, and the nominal service duration is decomposed according to the number of skill domains required by the task node to obtain the basic time consumption corresponding to each skill domain. For each required skill domain of the task node, the proficiency level of the technician in that skill domain and the required proficiency level of the task node in that skill domain are read, and the corresponding adjustment coefficient or adjustment rule is output by the service duration mapping relationship to correct the base consumption time, thereby obtaining the sub-service consumption time of that skill domain. The service duration mapping relationship is configured to make the sub-service consumption time of technicians with higher proficiency levels shorter when meeting the same required proficiency level. The time spent on each of the sub-services is aggregated across all skill domains involved in the task node to obtain the time taken for the technician to perform the experience-related service of the task node.

4. The experience-based field technician dispatching method according to claim 1, characterized in that, The feasibility constraints also include: Each accessed task node has a unique predecessor node and a unique successor node in the corresponding technician's work route to maintain the connectivity of the work route. Each technician can form at most one work route that starts from the network node and returns to the network node. The construction of the scheduling scheme includes maintaining auxiliary time information for each of the work routes. The auxiliary time information includes at least the earliest feasible service start time information and the latest allowed service start time information or time margin information along the work route. When performing cross-route movement, insertion, or replacement operations, the auxiliary time information is used to recursively calculate the affected section to determine whether the adjusted work route meets the task time window and the work time window without traversing the entire work route, and the auxiliary time information is updated synchronously when it is determined to be feasible.

5. The experience-based field technician dispatching method according to claim 1, characterized in that, The heuristic insertion generation initial scheduling scheme includes: Establish a set of unserved task nodes and initialize empty work routes for each technician; For each of the task nodes in the set of unserved task nodes, enumerate the candidate insertion positions for inserting the task node into the work route of any technician compatible with the task node, and calculate the change in the travel time of the work route before and after insertion for each candidate insertion position as the travel time increment. At the same time, verify the feasibility of the candidate insertion position in terms of time window and return point according to the feasibility constraints. Select the candidate insertion position with the smallest travel time increment from all feasible candidate insertion positions and perform the insertion, and remove the inserted task node from the set of unserved task nodes; The aforementioned enumeration, calculation, and insertion are performed repeatedly until there are no candidate insertion positions in the set of unserved task nodes that satisfy the feasibility constraints.

6. The experience-based field technician dispatching method according to claim 1, characterized in that, The local search includes a route structure optimization process to reduce the travel time of the work route. This process sequentially performs cross-route relocation, cross-route swap, intra-route relocation, intra-route swap, and arc set disconnection and reconnection operations on the candidate solutions according to a preset order. Specifically, the cross-route relocation operation removes an assigned task node from one work route and inserts it into the target position of another work route; the cross-route swap operation exchanges the positions of task nodes between different work routes; the intra-route relocation and swap operations change the order of task nodes within the same work route; and the arc set disconnection and reconnection operation selects a set of connecting arcs in the work route and changes the connection method between these arcs to rearrange the intermediate task node sequence. Furthermore, each operation adopts a first-time improvement criterion: when a feasible movement that reduces the total travel time and satisfies the feasibility constraints is found, the candidate solution is updated, and the search continues from the preset order of initial operations.

7. The experience-based field technician dispatching method according to claim 1, characterized in that, The local search also includes a benefit improvement process to enhance profitability, which includes insertion and replacement operations; wherein: The insertion operation is used to insert task nodes from the set of unserved task nodes into candidate positions on the work route, provided that the feasibility constraints are met. An insertion evaluation value is calculated for each candidate insertion scheme. This evaluation value is determined by the ratio of the nonlinearly amplified benefit of the task node to be inserted to the corresponding travel time increment. Furthermore, a disturbance coefficient randomly generated within a preset fluctuation range is introduced into the insertion evaluation value to change the ranking of the evaluation results in different iterations. Candidate insertion schemes are selected based on the insertion evaluation value, and insertion is performed until no feasible insertion is possible. The replacement operation is used to select the visited task node as the task node to be replaced, under the premise of satisfying the feasibility constraints, and to select the replacement task node from the set of unserved task nodes, so that the incremental benefit after replacement is positive and as large as possible. At the same time, the auxiliary time information is used to achieve rapid judgment in the replacement feasibility verification and replacement evaluation value calculation.

8. The experience-based field technician dispatching method according to claim 1, characterized in that, The determination and updating of the temperature parameter includes: during initialization, setting an initial value of the temperature parameter by logarithmic backpropagation based on the benefit level of the initial scheduling scheme, the preset degradation tolerance ratio, and the preset acceptance probability, so that when the benefit of a candidate scheme decreases relative to the initial scheduling scheme by the degradation tolerance ratio, it is still accepted with the acceptance probability; during the iteration process, the probability of acceptance is calculated by an exponential decay rule, and the acceptance probability decreases as the benefit difference between the candidate scheme and the current scheme increases, and a decision is made on whether to update the current scheme with the candidate scheme based on the comparison result of a random number and the acceptance probability; the disturbance includes selecting a random starting point for each of the job routes and continuously removing a number of assigned task nodes, and performing re-insertion processing on the unserved task nodes after removal to restore feasible benefits, while increasing, resetting, or limiting the removal scale according to a preset adjustment rule, and gradually reducing the temperature parameter according to a preset cooling rule; the termination condition includes the number of consecutive unimproved times reaching a preset upper limit and / or the number of iteration rounds reaching a preset upper limit.

9. An experience-based field technician dispatching system, the system comprising: The scheduling input construction module is used to acquire scheduling input data, which includes: a directed graph containing network nodes and multiple task nodes and the travel time between nodes, a set of available technical personnel, skill domain information, the proficiency level of each technical personnel in the skill domain information, the skill domain requirements and required proficiency levels of each task node, the revenue and task time window of each task node, the working time window of the network node, and the service duration mapping relationship used to determine service consumption time. The compatibility and service duration calculation module is used to generate a compatibility relationship between the technician and the task node based on the proficiency level and the required proficiency level, and to calculate the experience-related service duration of each compatibility combination in combination with the service duration mapping relationship. The feasibility constraint modeling module is used to construct a scheduling scheme under the condition of satisfying preset feasibility constraints. The feasibility constraints include at least: the unique allocation of the task node, the operation route of the technician starting from the network node and returning to the network node, and the service start time determined based on the travel time and the experience-related service duration to satisfy the task time window and the work time window. The heuristic initial solution generation module is used to generate an initial scheduling scheme using heuristic insertion. The heuristic insertion is used to select a feasible task node from the unserved task nodes and insert it into the work route of the technician compatible with the task node to reduce the travel time increment until no feasible insertion is available. An iterative local search optimization module is used to perform an iterative local search on the initial scheduling scheme. The iterative local search includes: generating a neighborhood scheme through local search; accepting the neighborhood scheme based on a temperature parameter when it is inferior to the current scheme; perturbing the current scheme when it is not accepted; and cooling the temperature parameter during the iteration process until the termination condition is met. The optimal solution output module is used to output the optimal scheduling scheme obtained during the iteration process. The optimal scheduling scheme is used to maximize the total revenue of the served task nodes while satisfying the feasibility constraints.

Images