Resource allocation and scheduling method based on state transition operator

Abstract

A resource allocation and scheduling method based on state transition operator, the method comprises the following steps: S1, constructing a state transition operator T according to the initial state of the water truck i and the final state of the water truck i; wherein i is a pointer without actual operation meaning; K is the total decision step number; S2, constructing a state transition operator branch corresponding to the decision action according to the state transition operator T and the decision action; the decision action comprises a work behavior, a driving behavior, a water supplementing behavior and a parking behavior; the state transition operator branch comprises time evolution, water quantity evolution, geographical position evolution and load rate evolution; S3, priority weight-based comprehensive scheduling cost target function F optimization. The method has the advantages of realizing efficient scheduling of the water truck.

Description

Technical Field

[0001] This invention relates to the fields of airport support vehicle scheduling and complex system modeling, and in particular to a resource allocation and scheduling method based on state transition operators. Background Technology

[0002] In airport operation support systems, water truck dispatching is a crucial link in ensuring normal flight turnaround and improving ground support efficiency, representing a typical application scenario for complex system resource allocation and scheduling. With the surge in airport flight throughput, water truck dispatching needs to balance complex requirements such as multi-task timing constraints, multi-vehicle resource coordination, and multi-geographic node linkage. It must also handle dynamic situations such as flight delays and temporary task assignments, placing extremely high demands on the accuracy, real-time performance, and robustness of dispatching modeling. In the field of complex system scheduling, accurately depicting the spatiotemporal relationships and dynamic evolution patterns of resources and tasks through modeling is a core prerequisite for achieving scheduling optimization. The completeness of related modeling technologies directly determines the feasibility, efficiency, and reliability of the scheduling scheme.

[0003] In existing technologies, time-space networks (TSNs) have become the mainstream modeling method for water truck scheduling optimization due to their ability to intuitively overlay geospatial and temporal information, clearly depicting the spatial location and task execution status of water trucks at different time points, and are widely used in practical engineering. However, TSN modeling requires the construction of a large number of nodes and arcs, resulting in high model complexity and computational overhead, making it difficult to adapt to the real-time requirements of lightweight scheduling systems. Therefore, engineers often choose not to explicitly construct the nodes and arcs of the TSN, but instead use state-transition graphs based on algebraic logic for equivalent modeling. These models abstract the vehicle scheduling process as a continuous evolution of system states, replacing the geometric description of time-space arcs with state vectors and transition rules. They offer advantages such as computational simplicity and adaptability to lightweight systems, and are gradually becoming an alternative solution in some scenarios.

[0004] While state transition graph (STP) modeling offers engineering advantages, current scheduling solutions based on this method still suffer from significant technical shortcomings. Existing STP modeling often focuses on single-dimensional state representation, failing to fully integrate multi-dimensional key information such as vehicle geographical location, water tank capacity, and task timing, resulting in incomplete state descriptions. Furthermore, the design of state transition rules is relatively simple, failing to adequately consider practical needs such as differences in task priorities and dynamic operating conditions, making it difficult to achieve precise matching of resources and tasks. These issues render existing STP models insufficient in scheduling accuracy and robustness, unable to meet the efficient scheduling requirements of high-throughput airport scenarios. Therefore, there is an urgent need to propose a STP scheduling method that balances multi-dimensional state representation with precise transition rules, improving the model's adaptability to complex scheduling scenarios. This is the core research and development objective of this invention. Summary of the Invention

[0005] The purpose of this invention is to provide a resource allocation and scheduling method based on state transition operators, which has the advantage of achieving efficient scheduling of water trucks.

[0006] To achieve the above objectives, the present invention provides a resource allocation and scheduling method based on a state transition operator. The method includes:

[0007] Step S1, based on the initial state of water truck i And the final state of the water truck i Construct a state transition operator T; where i is a descriptive term with no actual computational meaning; K is the total number of decision steps;

[0008] Step S2: Based on the state transition operator T and the decision action, construct the state transition operator branch corresponding to the decision action; the decision action includes: operation behavior, driving behavior, water replenishment behavior, and dwell behavior; the state transition operator branch includes: time evolution, water volume evolution, geographical location evolution, and load rate evolution.

[0009] Step S3: Optimize the overall scheduling cost objective function F based on priority weights.

[0010] Preferably, the method includes: step S1 includes:

[0011] Step S11: Obtain the initial state of water truck i The initial state Includes: initial location set as a fixed stop for the airport water truck, initial water volume at full tank, initial timestamp as... The initial load rate is ;

[0012] Step S12: Obtain the final state of water truck i. ;

[0013] Among them, the final state Includes the location, water volume, timestamp, and load rate of water truck i after the decision in step K;

[0014] Step S13: Construct the state transition operator T for the water truck, satisfying... ;

[0015] in, For the water truck i in the first The decision action of step, i is a reference and has no actual computational meaning; ; For work-related behaviors, For driving behavior, For the purpose of replenishing water, For the purpose of staying; For the water truck i in the first The next state vector of the step.

[0016] Preferably, the method includes:

[0017] In step S2, the state transition operator branch corresponding to the decision action includes the job behavior state transition operator branch. Driving behavior state transition operator branch Water replenishment behavior state transition operator branch and the state transition operator branch for dwelling behavior ;

[0018] In step S2, the water volume evolution includes: the water volume evolution corresponding to the operation behavior and the water volume evolution corresponding to the water replenishment behavior.

[0019] Preferably, in step S2, the water volume evolution corresponding to the operational behavior includes:

[0020] Step Q1: Data processing, retrieve water consumption data from the past three operations of the water cleaning truck. , , Current water tank balance Current task water requirements Current timestamp and the estimated execution time of the task By using an outlier removal algorithm, abnormal consumption data caused by equipment failure or human error in historical data is removed to ensure the accuracy of input data.

[0021] Step Q2, weighting coefficient calculation, uses an adaptive weighting algorithm to dynamically calculate the weight of historical data based on the time difference between the current timestamp and the historical task timestamp, and the deviation between the current task water demand and the historical task water demand. and current data weights The weights satisfy The specific calculation formula is as follows: , ;

[0022] in, The average timestamp of historical tasks. This represents the average water demand for historical operations.

[0023] Step Q3, calculation of operating condition correction factor, introducing dynamic operating condition correction factor. It is used to correct the impact of dynamic scenarios such as flight delays and temporary additional tasks on water consumption; The calculation is based on the urgency of the current task and real-time operational fluctuations. The calculation formula is: ;

[0024] in, Let be the delay penalty coefficient for task with priority j. This represents real-time operational fluctuations, measured by the current airport flight punctuality rate. Calculations show that The lower the flight punctuality rate, the greater the fluctuation in operating conditions. The larger the value, the more dynamically it can adapt to the differences in water consumption during sudden events;

[0025] Step Q4: Final calculation of water volume evolution. Combining the preprocessed data, weighting coefficients, and correction factors mentioned above, calculate the remaining water tank capacity after the operation. The final result of water volume evolution is calculated as follows: Simultaneously, the calculation result is compared with the preset water safety redundancy. Make a comparison and judgment, if If the water volume evolution is valid, then subsequent state updates will be performed; if If the current water truck is unable to complete the task, it will be switched to another water truck for dispatch.

[0026] Preferably, in step Q1, the calculation logic of the outlier removal algorithm is as follows: first calculate the mean of historical water consumption data. Then calculate the standard deviation. If a certain historical data satisfies If i=1,2,3, it is considered an outlier and is removed. The mean μ is then recalculated using the remaining data as a historical reference benchmark.

[0027] Preferably, in step S2, the water volume evolution corresponding to the water replenishment behavior includes:

[0028] Step W1, data retrieval, retrieve the water replenishment rate of the water replenishment point. The water tank of the water truck is full of water volume. Current water tank balance and the timestamp before hydration ;

[0029] Step W2: Calculate the water replenishment duration and water volume evolution, including the water replenishment duration. and the remaining water level in the tank after refilling. The formula for calculating the water replenishment time is as follows: The formula for calculating the remaining water level in the tank after replenishment is: .

[0030] Preferably, step S3 includes: Step S31, based on the core data of water evolution, performing least squares fitting on the relevant data to eliminate random errors in the water evolution calculation process, ensuring that the objective function F can accurately reflect the correlation between scheduling cost and water evolution state, the formula is: ;

[0031] The fitted data includes: weighting coefficients in water evolution. , Operating condition correction factor The remaining water level in the tank after operation / refilling , Homework duration Hydration duration Comprehensive cost item Delay penalty costs And the state transition operator T;

[0032] Step S32: Solve for the optimal value of the objective function F. The minimum value of F is found by using an improved particle swarm optimization algorithm. This algorithm is adapted to the dynamic characteristics of water volume evolution data. The specific solution logic is as follows: take the fitted water volume evolution parameters as the initial position of the particles, and use the calculated value of F as the fitness function. By iteratively updating the particle velocity and position, the water volume evolution path and operation / replenishment sequence of each water truck are continuously optimized. Finally, the particle parameter combination corresponding to the minimum value of F is obtained. This parameter combination is the optimal water volume evolution parameters and the optimal scheduling scheme of the water truck.

[0033] Preferably, the fitting process in step S31 specifically involves: combining the theoretically calculated values ​​of water volume evolution with the measured values ​​in the actual scheduling scenario. Using the sample as an example, the sum of squared deviations between the theoretical and measured values ​​is minimized using the least squares fitting algorithm. Correcting the weighting coefficients and operating condition correction factor The calculation parameters are determined so that the fitted data can accurately match the actual scheduling conditions; based on the fitted data, a comprehensive scheduling cost objective function F is constructed.

[0034] Preferably, in step S3, ;in, , , as well as as follows:

[0035] Time cost Operation duration corresponding to water volume evolution Hydration duration The formula for calculating the fitted data is as follows: , The time cost coefficient is used to ensure that the fitted data accurately reflects the time consumption during the water volume evolution process.

[0036] Energy consumption cost : The remaining capacity of the water tank after fitting , The derivation is as follows: , As the energy consumption cost coefficient, the more accurate the water volume evolution, the more realistic the energy consumption cost calculation.

[0037] Resource idle cost : Redundant water volume that has not been utilized in the evolution of water volume ( According to the fitted data, the more redundant water, the higher the idle cost.

[0038] Delay penalty costs : Operating condition correction factor in water quantity evolution The linkage and fitting of the data ensure that the delay penalty can match the dynamic operating conditions adapted to the water volume evolution.

[0039] In summary, compared with the prior art, the resource allocation and scheduling method based on state transition operators provided by this invention has the following beneficial effects:

[0040] First, this invention breaks through the dependence of traditional spatiotemporal networks on the physical definition of "arc segments". It integrates the spatial, temporal, resource, and load attributes of vehicles through multi-dimensional state vectors, and uses state transition operators to characterize the core logic of the scheduling process, thus achieving an equivalent characterization of complex scheduling logic. Moreover, the modeling process is simpler and more in line with engineering implementation requirements.

[0041] Second, the state transition logic of this invention adopts an algebraic operator mapping form, which is easier to combine with modern artificial intelligence algorithms such as reinforcement learning (RL) and deep neural networks (DNN). It can directly use the state vector as the input feature of the intelligent algorithm and the state transition operator as the output mapping of the intelligent decision. It reserves a flexible technical interface for the future intelligent upgrade of the scheduling system and expands the application scenarios and development space of the technical solution. Attached Figure Description

[0042] Figure 1 This is a flowchart of a resource allocation and scheduling method based on state transition operators proposed in this invention. Detailed Implementation

[0043] The following will be combined with the appendix in the embodiments of the present invention. Figure 1 The technical solutions, structural features, objectives and effects achieved in the embodiments of the present invention will be described in detail.

[0044] It should be noted that the accompanying drawings are in a very simplified form and use non-precise proportions. They are only used to facilitate and clarify the purpose of illustrating the embodiments of the present invention, and are not intended to limit the implementation conditions of the present invention. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportional relationship, or adjustments to the size should still fall within the scope of the technical content disclosed in the present invention, provided that they do not affect the effects and objectives that the present invention can produce.

[0045] It should be noted that, in this invention, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only the expressly listed elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0046] like Figure 1 As shown, the purpose of this invention is to propose a resource allocation and scheduling method based on state transition operators. By defining a vehicle state vector containing multi-dimensional information such as geography, time, water volume, and load, and establishing state transition rules linked to task priority, efficient scheduling can be achieved without relying on explicit spatiotemporal network topology.

[0047] The method includes:

[0048] Step S1, based on the initial state of water truck i And the final state of the water truck i Construct a state transition operator T; where i is a descriptive term with no actual computational meaning; K is the total number of decision steps;

[0049] Step S2: Based on the state transition operator T and the decision action, construct the state transition operator branch corresponding to the decision action; the decision action includes: operation behavior, driving behavior, water replenishment behavior and dwell behavior; the state transition operator branch includes: time evolution, water volume evolution, geographical location evolution and load rate evolution.

[0050] The key point of this invention is water volume evolution.

[0051] Step S3: Optimize the overall scheduling cost objective function F based on priority weights.

[0052] Specifically, step S1 includes:

[0053] Step S11: Obtain the initial state of water truck i The initial state Includes: initial location set as a fixed stop for the airport water truck, initial water volume at full tank, initial timestamp as... The initial load rate is ;

[0054] Step S12: Obtain the final state of water truck i. ;

[0055] Among them, the final state Includes the location, water volume, timestamp, and load rate of water truck i after the decision in step K;

[0056] Step S13: Construct the state transition operator T for the water truck, satisfying... ;

[0057] in, For the water truck i in the first The decision action of step, i is a reference and has no actual computational meaning; ; For work-related behaviors, For driving behavior, For the purpose of replenishing water, For the purpose of staying; For the water truck i in the first The next state vector of the step.

[0058] It should be clarified here that the state transition operator T of the present invention differs from the single-dimensional transition rules in the prior art. It can integrate multi-dimensional data such as water volume and time, providing core support for subsequent water volume evolution algorithms and reflecting fundamental innovation.

[0059] Specifically, in step S2, the state transition operator branch corresponding to the decision action includes the job behavior state transition operator branch. Driving behavior state transition operator branch Water replenishment behavior state transition operator branch and the state transition operator branch for dwelling behavior .

[0060] The state transition operator branch corresponding to the decision action is centered on the water volume evolution branch, while the other branches of time evolution, geographical location evolution, and load rate evolution serve only as auxiliary functions to complete the state update in conjunction with the water volume evolution (details will not be elaborated here). Among them, the water volume evolution branch corresponds to the core state transition of the operation behavior and water replenishment behavior. An adaptive weighted recursive algorithm is used to realize the dynamic evolution of water volume data. This algorithm is one of the core innovations of this invention. Unlike the simple addition and subtraction formulas in the existing technology, it improves the accuracy of water volume evolution by integrating historical water volume data and real-time operating condition data for weighted calculation, and is suitable for the dynamic scheduling scenario of water trucks.

[0061] This step focuses on the core algorithm design for water volume evolution, clarifying the principle, data utilization method, and detailed calculation process of the adaptive weighted recursive algorithm. It highlights the difference from existing simple judgments and formulas, demonstrating technological innovation. In existing technologies, water volume calculation for water trucks only uses simple addition and subtraction operations such as "current water volume - task water requirement" and "current water volume + replenishment water volume," without considering factors such as historical data deviations and dynamic operating condition fluctuations, resulting in large water volume calculation errors and making it unsuitable for scenarios such as flight delays and temporary tasks. The adaptive weighted recursive algorithm of this invention solves the above-mentioned technical defects through dynamic weighting coefficient adjustment and operating condition correction. The specific core logic is: based on the historical water consumption data of the water truck, the current water tank balance data, and the real-time task requirement data, the water tank balance is accurately calculated through dynamic weighting coefficient adjustment. At the same time, an operating condition correction factor is introduced to cope with dynamic scenarios such as flight delays and temporary tasks, avoiding the defects of the "one-size-fits-all" simple formula in existing technologies.

[0062] In a specific embodiment, step S2, the water volume evolution includes:

[0063] The water volume evolution corresponding to operational activities and the water volume evolution corresponding to replenishment activities will be explained one by one below.

[0064] The water volume evolution corresponding to the aforementioned operational behavior includes:

[0065] Step Q1: Data processing, retrieve water consumption data from the past three operations of the water cleaning truck. , , Current water tank balance Current task water requirements Current timestamp and the estimated execution time of the task By using an outlier removal algorithm (emphasizing the 3σ criterion), abnormal consumption data caused by equipment failure or human error in historical data is removed, ensuring the accuracy of the input data.

[0066] Specifically, the outlier removal algorithm works as follows: first, calculate the mean of historical water consumption data. Then calculate the standard deviation. If a certain historical data satisfies If i=1,2,3, it is considered an outlier and is removed. The mean μ is then recalculated using the remaining data as a historical reference benchmark.

[0067] Among them, the 3σ criterion can effectively eliminate extreme abnormal data, ensure the reliability of historical water volume data, and provide accurate input for subsequent weighted calculations. This design is different from the shortcomings of existing technologies that do not handle abnormal data.

[0068] Step Q2, weighting coefficient calculation, uses an adaptive weighting algorithm to dynamically calculate the weight of historical data based on the time difference between the current timestamp and the historical task timestamp, and the deviation between the current task water demand and the historical task water demand. and current data weights The weights satisfy The specific calculation formula is as follows: , ;

[0069] in, The average timestamp of historical tasks. The average water demand of historical operations is used; the dynamic weighting logic gives higher weight to historical data that is more recent and has smaller deviations in water demand, which solves the problem of insufficient accuracy caused by fixed weight or no weight calculation in the existing technology, and is one of the key innovations of this invention.

[0070] Step Q3, calculation of operating condition correction factor, introducing dynamic operating condition correction factor. It is used to correct the impact of dynamic scenarios such as flight delays and temporary additional tasks on water consumption; The calculation is based on the urgency of the current task and real-time operational fluctuations. The calculation formula is: ;

[0071] in, Let be the delay penalty coefficient for task with priority j. This represents real-time operational fluctuations, measured by the current airport flight punctuality rate. Calculations show that The lower the flight punctuality rate, the greater the fluctuation in operating conditions. The larger the value, the more dynamically it can adapt to the differences in water consumption in sudden scenarios, filling the technological gap in existing technologies that do not consider dynamic working conditions.

[0072] Step Q4: Final calculation of water volume evolution. Combining the preprocessed data, weighting coefficients, and correction factors mentioned above, calculate the remaining water tank capacity after the operation. The final result of water volume evolution is calculated as follows: Simultaneously, the calculation result is compared with the preset water safety redundancy. Make a comparison and judgment, if If the water volume evolution is valid, then subsequent state updates will be performed; if If the current water truck is unable to complete the task, it will be switched to another water truck for dispatch.

[0073] This formula integrates historical, real-time, and operational data to achieve multi-dimensional linkage calculations, which differs from the simple addition and subtraction operations of existing technologies. It is also deeply bound to the state transition operator T to ensure the accuracy of state transitions.

[0074] The water volume evolution corresponding to the water replenishment behavior includes:

[0075] Step W1, data retrieval, retrieve the water replenishment rate of the water replenishment point. The water tank of the water truck is full of water volume. Current water tank balance (as mentioned above) (Linkage to enable data reuse) and timestamps before water replenishment .

[0076] This step, along with the aforementioned data reuse design, simplifies the calculation process and avoids the tediousness of repeatedly collecting data in existing technologies, representing an optimization and innovation in engineering implementation.

[0077] Step W2: Calculate the water replenishment duration and water volume evolution, including the water replenishment duration. and the remaining water level in the tank after refilling. The formula for calculating the water replenishment time is as follows: (Introducing operating condition correction factors) (To ensure adaptability to dynamic operating conditions); the formula for calculating the remaining water level in the tank after replenishment is: .

[0078] Note: The formula reuses the aforementioned core parameters to achieve algorithm consistency, while also modifying the water replenishment time to adapt to dynamic working conditions, which is different from the simple water replenishment logic of existing technology that "stops when the tank is full".

[0079] The temporal evolution, geographic location evolution, and load rate evolution are all implemented in conjunction with the results of the aforementioned water volume evolution algorithm, and will not be elaborated upon here; among them, the timestamp update is combined with the operation duration during the water volume evolution process. Hydration duration The calculations show that geographic location evolution only updates location identifiers, while load factor evolution is based on water volume evolution results. With full tank of water Calculation of the ratio ( Each evolutionary branch serves the core of water quantity evolution, avoiding insufficient creativity caused by the superposition of multiple evolutions, while ensuring the integrity of the state transition process.

[0080] Specifically, step S3 includes:

[0081] Step S31: Based on the core data of water volume evolution, perform least squares fitting on the relevant data to eliminate random errors in the water volume evolution calculation process (such as historical data sampling errors and calculation deviations caused by operating condition fluctuations), ensuring that the objective function F can accurately reflect the correlation between scheduling costs and water volume evolution status. The formula is as follows: ;

[0082] The fitted data includes: weighting coefficients in water evolution. , Operating condition correction factor The remaining water level in the tank after operation / refilling , Homework duration Hydration duration Comprehensive cost item And the state transition operator T.

[0083] Least squares fitting enables deep linkage between water evolution data and objective function, solving the defect of the objective function being disconnected from the state transition rules in existing technologies and improving the accuracy of scheduling cost calculation.

[0084] The fitting process specifically involves using theoretically calculated values ​​of water evolution (such as...) Theoretical values ​​and measured values ​​in actual scheduling scenarios Using the sample as an example, the sum of squared deviations between the theoretical and measured values ​​is minimized using the least squares fitting algorithm. Correcting the weighting coefficients and operating condition correction factor The calculation parameters are determined so that the fitted data can accurately match the actual scheduling conditions; based on the fitted data, a comprehensive scheduling cost objective function F is constructed.

[0085] Each cost term in the objective function is linked to water volume evolution data, forming a closed-loop technology of "state transition → water volume evolution → cost optimization," ensuring the practicality and accuracy of the scheduling scheme. Among these, the comprehensive cost term... The revised complete formula, which is deeply tied to water volume evolution, is as follows: Each cost item is derived from the fitted water volume evolution data, achieving a one-to-one correspondence with the state transition operator branches, further strengthening the technical closed loop.

[0086] in, , as well as as follows:

[0087] 1. Time cost Operation duration corresponding to water volume evolution Hydration duration The formula for calculating the fitted data is as follows: , The time cost coefficient is used to ensure that the fitted data accurately reflects the time consumption during the water volume evolution process.

[0088] 2. Energy consumption cost : The remaining capacity of the water tank after fitting , The derivation is as follows: , As the energy consumption cost coefficient, the more accurate the water volume evolution, the more realistic the energy consumption cost calculation.

[0089] 3. Resource idleness costs : Redundant water volume that has not been utilized in the evolution of water volume ( According to the fitted data, the more redundant water, the higher the idle cost.

[0090] 4. Costs of penalties for delays : Operating condition correction factor in water quantity evolution Linkage (as already stated above) The fitted data ensures that the delay penalty can match the dynamic operating conditions adapted to the water volume evolution.

[0091] Each cost item is deeply linked to water volume evolution, ensuring that the scheduling optimization objectives and state transition processes are accurately aligned, thereby improving the feasibility of the solution.

[0092] Step S32: Solve for the optimal value of the objective function F. An improved particle swarm optimization algorithm is used to find the minimum value of F (the smaller F is, the lower the scheduling cost and the better the solution). This algorithm is adapted to the dynamic characteristics of water evolution data and differs from existing simple solution algorithms. The specific solution logic is as follows: using the fitted water evolution parameters (… , , Using F as the initial position of the particle and F as the fitness function, the particle velocity and position are updated iteratively to continuously optimize the water volume evolution path and operation / replenishment sequence of each water truck. Finally, the particle parameter combination corresponding to the minimum value of F is obtained. This parameter combination is the optimal water volume evolution parameter, which corresponds to the best scheduling scheme of the water truck (including optimal operation allocation, water replenishment timing, and vehicle dispatching order).

[0093] The improved particle swarm optimization algorithm introduces water quantity evolution constraints to solve the problems of existing solution algorithms not adapting to dynamic working conditions and having low solution accuracy, ensuring that the scheduling scheme is optimal and feasible.

[0094] In addition, the iteration termination condition is: the deviation of the F value in 10 consecutive iterations is less than 10−5, to ensure the stability and accuracy of the solution. The optimal solution obtained is essentially to achieve the optimal mapping of the state transition operator T by optimizing the core parameters of water evolution, forming a complete technical closed loop with the state transition operator branch mentioned above, further highlighting the inventiveness of the present invention and distinguishing it from the defect of the objective function being disconnected from the state transition rule in the prior art.

[0095] Although the present invention has been described in detail through the preferred embodiments above, it should be understood that the above description should not be considered as a limitation of the present invention. Various modifications and substitutions to the present invention will be apparent to those skilled in the art after reading the above description. Therefore, the scope of protection of the present invention should be defined by the appended claims.

Claims

1. A resource allocation and scheduling method based on a state transition operator, characterized in that, The method includes: Step S1, based on the initial state of water truck i And the final state of the water truck i Construct a state transition operator T; where i is a descriptive term with no actual computational meaning; K is the total number of decision steps; Step S2: Based on the state transition operator T and the decision action, construct the state transition operator branch corresponding to the decision action; the decision action includes: operation behavior, driving behavior, water replenishment behavior, and dwell behavior; the state transition operator branch includes: time evolution, water volume evolution, geographical location evolution, and load rate evolution. Step S3: Optimize the overall scheduling cost objective function F based on priority weights.

2. The resource allocation and scheduling method based on state transition operators according to claim 1, characterized in that, The method includes: Step S1 includes: Step S11: Obtain the initial state of water truck i The initial state Includes: initial location set as a fixed stop for the airport water truck, initial water volume at full tank, initial timestamp as... The initial load rate is ; Step S12: Obtain the final state of water truck i. ; Among them, the final state Includes the location, water volume, timestamp, and load rate of water truck i after the decision in step K; Step S13: Construct the state transition operator T for the water truck, satisfying... ; in, For the water truck i in the first The decision action of step, i is a reference and has no actual computational meaning; ; For work-related behaviors, For driving behavior, For the purpose of replenishing water, For the purpose of staying; For the water truck i in the first The next state vector of the step.

3. The resource allocation and scheduling method based on a state transition operator according to claim 2, characterized in that, The method includes: In step S2, the state transition operator branch corresponding to the decision action includes the job behavior state transition operator branch. Driving behavior state transition operator branch Water replenishment behavior state transition operator branch and the state transition operator branch for dwelling behavior ; In step S2, the water volume evolution includes: the water volume evolution corresponding to the operation behavior and the water volume evolution corresponding to the water replenishment behavior.

4. The resource allocation and scheduling method based on a state transition operator according to claim 3, characterized in that, In step S2, the water volume evolution corresponding to the operation includes: Step Q1: Data processing, retrieve water consumption data from the past three operations of the water cleaning truck. , , Current water tank balance Current task water requirements Current timestamp and the estimated execution time of the task By using an outlier removal algorithm, abnormal consumption data caused by equipment failure or human error in historical data is removed to ensure the accuracy of input data. Step Q2, weighting coefficient calculation, uses an adaptive weighting algorithm to dynamically calculate the weight of historical data based on the time difference between the current timestamp and the historical task timestamp, and the deviation between the current task water demand and the historical task water demand. and current data weights The weights satisfy The specific calculation formula is as follows: , ; in, The average timestamp of historical tasks. This represents the average water demand for historical operations. Step Q3, calculation of operating condition correction factor, introducing dynamic operating condition correction factor. It is used to correct the impact of dynamic scenarios such as flight delays and temporary additional tasks on water consumption; The calculation is based on the urgency of the current task and real-time operational fluctuations. The calculation formula is: ; in, Let be the delay penalty coefficient for task with priority j. This represents real-time operational fluctuations, measured by the current airport flight punctuality rate. Calculations show that The lower the flight punctuality rate, the greater the fluctuation in operating conditions. The larger the value, the more dynamically it can adapt to the differences in water consumption during sudden events; Step Q4: Final calculation of water volume evolution. Combining the preprocessed data, weighting coefficients, and correction factors mentioned above, calculate the remaining water tank capacity after the operation. The final result of water volume evolution is calculated as follows: Simultaneously, the calculation result is compared with the preset water safety redundancy. Make a comparison and judgment, if If the water volume evolution is valid, then subsequent state updates will be performed; if If the current water truck is unable to complete the task, it will be switched to another water truck for dispatch.

5. The resource allocation and scheduling method based on a state transition operator according to claim 4, characterized in that, In step Q1, the calculation logic of the outlier removal algorithm is as follows: first calculate the mean of historical water consumption data. Then calculate the standard deviation. ; If a certain historical data satisfies If i=1,2,3, it is considered an outlier and is removed. The mean μ is then recalculated using the remaining data as a historical reference benchmark.

6. The resource allocation and scheduling method based on a state transition operator according to claim 5, characterized in that, In step S2, the water volume evolution corresponding to the water replenishment behavior includes: Step W1, data retrieval, retrieve the water replenishment rate of the water replenishment point. The water tank of the water truck is full of water volume. Current water tank balance and the timestamp before hydration ; Step W2: Calculate the water replenishment duration and water volume evolution, including the water replenishment duration. and the remaining water level in the tank after refilling. The formula for calculating the water replenishment time is as follows: The formula for calculating the remaining water level in the tank after replenishment is: .

7. The resource allocation and scheduling method based on a state transition operator according to claim 6, characterized in that, Step S3 includes: Step S31: Based on the core data of water volume evolution, perform least squares fitting on the relevant data to eliminate random errors in the water volume evolution calculation process, ensuring that the objective function F accurately reflects the correlation between scheduling cost and water volume evolution state. The formula is as follows: ; The fitted data includes: weighting coefficients in water evolution. , Operating condition correction factor The remaining water level in the tank after operation / refilling , Homework duration Hydration duration Comprehensive cost item Delay penalty costs And the state transition operator T; Step S32: Solve for the optimal value of the objective function F. The minimum value of F is found by using an improved particle swarm optimization algorithm. This algorithm is adapted to the dynamic characteristics of water volume evolution data. The specific solution logic is as follows: take the fitted water volume evolution parameters as the initial position of the particles, and use the calculated value of F as the fitness function. By iteratively updating the particle velocity and position, the water volume evolution path and operation / replenishment sequence of each water truck are continuously optimized. Finally, the particle parameter combination corresponding to the minimum value of F is obtained. This parameter combination is the optimal water volume evolution parameters and the optimal scheduling scheme of the water truck.

8. A resource allocation and scheduling method based on a state transition operator according to claim 7, characterized in that, The fitting process in step S31 specifically involves: comparing the theoretically calculated values ​​of water volume evolution with the measured values ​​in the actual scheduling scenario. Using the sample as an example, the sum of squared deviations between the theoretical and measured values ​​is minimized using the least squares fitting algorithm. Correcting the weighting coefficients and operating condition correction factor The calculation parameters are determined so that the fitted data can accurately match the actual scheduling conditions; based on the fitted data, a comprehensive scheduling cost objective function F is constructed.

9. A resource allocation and scheduling method based on a state transition operator according to claim 8, characterized in that, In step S3 ;in, , , as well as as follows: Time cost Operation duration corresponding to water volume evolution Hydration duration The formula for calculating the fitted data is as follows: , The time cost coefficient is used to ensure that the fitted data accurately reflects the time consumption during the water volume evolution process. Energy consumption cost : The remaining capacity of the water tank after fitting , The derivation is as follows: , As the energy consumption cost coefficient, the more accurate the water volume evolution, the more realistic the energy consumption cost calculation. Resource idle cost : Redundant water volume that has not been utilized in the evolution of water volume ( According to the fitted data, the more redundant water, the higher the idle cost. Delay penalty costs : Operating condition correction factor in water quantity evolution The linkage and fitting of the data ensure that the delay penalty can match the dynamic operating conditions adapted to the water volume evolution.

Images