Scheduling method, device and equipment for multiple crown blocks on same rail and storage medium
Through the combination of deep Q network and genetic algorithm, the optimal homo-rail multi-day vehicle scheduling solution is generated, which solves the problem that traditional scheduling technology cannot achieve multi-objective and multi-level optimization, improves the efficiency of the warehouse and considers environmental impact.
Patent Information
- Application Number
- CN202411115638.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-12-06
AI Technical Summary
Traditional homo-rail multi-day vehicle scheduling technology cannot achieve multi-target and multi-level optimal scheduling solutions, resulting in the overall inefficiency of the warehouse factory and the inability to effectively handle emergency tasks.
Deep Q network algorithm and genetic algorithm are used to combine carbon emissions, total completion time and total operating distance to generate initial populations, and the optimal scheduling scheme is generated through iterative updates, taking into account environmental impact and resource consumption, achieving high efficiency, low loss and environmentally friendly Scheduling.
Through multi-objective and multi-level optimization, the optimal scheduling solution is obtained, the overall efficiency of the warehouse factory is improved, the problem of sky-car conflict is solved, and carbon emission factors are introduced during the scheduling process, achieving the goal of high efficiency, low loss and environmental protection.
Smart Images

Figure CN119090197A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of unmanned overhead crane dispatching, and in particular to a dispatching method, device, equipment and storage medium for multiple overhead cranes on the same track. Background Art
[0002] With the development of industrial automation, the demand for logistics and production efficiency continues to grow. The same-track multi-crane scheduling technology is widely used in industrial production, railway transportation, port logistics and other fields to meet large-scale and high-efficiency logistics and production needs.
[0003] Traditional priority assessment relies on manual experience and cannot interact with the environment. It cannot assign priorities to goods according to demand and is a static scheduling method that cannot solve emergency tasks. At the same time, the scheduling plan is not optimal, which affects the overall efficiency of the warehouse. With the continuous development of artificial intelligence technology, the scheduling technology of multiple overhead cranes on the same track is also developing towards intelligence. Through integrated scheduling algorithms, the movement status of the overhead cranes can be monitored and scheduled in real time. However, the goals of these intelligent scheduling algorithms are relatively single, and it is impossible to obtain the optimal scheduling plan through multiple goals and multiple levels.
[0004] Therefore, there is an urgent need for a method for scheduling multiple overhead travelling vehicles on the same track, which can optimize the scheduling scheme with multiple objectives and multiple levels to obtain the optimal scheduling scheme. Summary of the invention
[0005] In view of this, it is necessary to provide a method, device, equipment and storage medium for scheduling multiple overhead travelling vehicles on the same track, which can optimize the scheduling scheme with multiple objectives and multiple levels to obtain the optimal scheduling scheme.
[0006] In order to solve the above technical problems, on the one hand, the present invention provides a method for dispatching multiple overhead travelling vehicles on the same track, comprising: Determine the priority of goods to be distributed; Generate an initial population according to the priority of the goods to be allocated, determine the fitness value of each chromosome in the initial population according to the initial population, the fitness value is determined by three factors: the total completion time, the total running distance and the carbon emission of each overhead crane in the chromosome, and determine the parent chromosome based on the fitness value of the chromosome; A new population is generated according to the parent chromosome, the new population is iteratively updated to generate a final population, and an optimal scheduling scheme for multi-SkyTrains on the same track is generated according to the fitness values of the chromosomes in the final population.
[0007] In a possible implementation, determining the priority of goods to be allocated includes: Obtain historical task data, current location of the overhead crane, assigned tasks to be completed, and material requirements; Based on the deep Q network algorithm, build the initial model; Determine the current environmental status based on the current position of the overhead crane, the assigned tasks to be completed, and the material requirements; Taking the type of goods to be allocated and the quality of goods to be allocated as input data, training an initial model according to the historical task data, the current environment state and the input data, generating a priority value model, and determining the priority value of the goods to be allocated; Based on the greedy algorithm, the priority value model is trained according to the priority value of the goods to be allocated, a priority division model is generated, and the priority of the goods to be allocated is determined.
[0008] In a possible implementation, generating an initial population according to the priorities of the goods to be allocated includes: Based on the priority division model, a scheduling model is constructed. Based on the genetic algorithm, a gene is assigned to each cargo according to the priority of the cargo to be assigned and the current environmental status. The value of the gene is set as the crane number, and multiple genes are generated. Generate multiple scheduling schemes based on the multiple genes, each scheduling scheme constituting a chromosome; The plurality of chromosomes constitute an initial population.
[0009] In a possible implementation, determining the fitness value of each chromosome in the initial population according to the initial population includes: The weighted sum of each crane in the scheduling plan is calculated based on the total completion time of each crane, the total running distance of the crane, and the carbon emissions of the crane. The weighted sum calculation formula of the crane is: , in, is the weighted sum of the i-th overhead travelling crane, is the total completion time of the i-th overhead crane to complete all tasks, is the total running distance of the ith crane to complete all tasks, The carbon emissions of the ith overhead crane completing all tasks, is the weight factor between distance and completion time, is the weight factor of carbon emissions; According to the weighted sum of the overhead crane, the fitness function of the chromosome is established, and the fitness function of the chromosome is: , in, is the fitness value of the chromosome, n is the total number of cranes in the chromosome, is the weighted sum of the nth overhead travelling crane; According to the fitness function of the chromosome, the maximum weighted sum of each overhead travelling vehicle is selected as the fitness value of the chromosome.
[0010] In a possible implementation, determining the parent chromosome based on the fitness value includes: The chromosomes whose fitness values are less than the preset fitness threshold are screened out and determined as parent chromosomes.
[0011] In a possible implementation, a new population is generated according to the parent chromosome, the new population is iteratively updated to generate a final population, and an optimal scheduling scheme is generated according to the fitness values of the chromosomes in the final population, including: Perform genetic crossover and mutation operations on the parent chromosome to generate a new population; Iteratively training and updating the scheduling scheme model using the new population as training data, recalculating the fitness value of each chromosome in the new population, determining the parent chromosome of the new population, performing crossover and mutation operations on the parent of the new population until a preset termination condition is reached to generate a final population; The chromosome with the smallest fitness value is selected from the chromosomes of the final population to generate an optimal scheduling solution.
[0012] In a possible implementation, after generating the optimal scheduling solution, the method further includes: The fitness value of the optimal scheduling solution is calculated to be inverse, and a reward return of the optimal scheduling solution is generated; Based on the time-domain difference method, updating the parameters of the scheduling model according to the reward return; The scheduling information of the overhead crane is updated according to the optimal scheduling scheme, and the optimal scheduling scheme is used as historical task data to update the database.
[0013] In a second aspect, the present invention further provides a dispatching device for multiple overhead travelling vehicles on the same track, comprising: A priority determination module, used to determine the priority of goods to be allocated; A parent chromosome determination module is used to generate an initial population according to the priority of the goods to be allocated, and determine the fitness value of each chromosome in the initial population according to the initial population, wherein the fitness value is determined by three factors: the total completion time, the total running distance, and the carbon emission of each overhead crane in the chromosome, and the parent chromosome is determined based on the fitness value of the chromosome; The optimal scheduling scheme generating module is used to generate a new population according to the parent chromosome, iteratively update the new population to generate a final population, and generate an optimal scheduling scheme for the multi-sky vehicles on the same track according to the fitness values of the chromosomes in the final population.
[0014] In a third aspect, the present invention further provides a device, comprising a memory and a processor, wherein the memory is used to store programs and data; the processor is coupled to the memory and is used to execute the program stored in the memory to implement the same-track multi-overhead vehicle scheduling method as described above.
[0015] In a fourth aspect, the present invention further provides a computer storage medium for storing a computer-readable program or instruction, which, when executed by a processor, can implement the same-track multi-overhead vehicle scheduling as described above.
[0016] The beneficial effects of the present invention are as follows: first, the present invention determines the priority of the goods to be allocated, and determining the priority of the goods to be allocated can solve the overhead crane conflict problem; then, based on the genetic algorithm, an initial population is generated according to the priority of the goods to be allocated, and the fitness value of the chromosome of the initial population is calculated according to the total completion time, total running distance and carbon emissions of the overhead crane, and the parent chromosome is determined; finally, the parent chromosome is iteratively updated to generate a final population, and according to the fitness value of the chromosome, the chromosome with the smallest fitness value is screened out in the final population as the optimal scheduling scheme. The present invention first determines the priority of the goods to be allocated, and the scheme is fitness screened according to the three objectives of the completion time, running distance and carbon emissions of the overhead crane, and cross-mutation and iterative update are performed at the same time, so as to achieve multi-objective and multi-level optimization and screening of the scheduling scheme, and obtain the optimal scheduling scheme. In the process of determining the scheduling scheme, the carbon emission factor is introduced, and the total completion time and total running distance of the overhead crane are combined, which not only considers the efficiency of task completion and resource consumption, but also considers the impact of the environment, and achieves the optimization and screening of the scheduling scheme with high efficiency, low loss and environmental protection. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.
[0018] Figure 1 A schematic flow chart of an embodiment of the method for dispatching multiple overhead travelling vehicles on the same track provided by the present invention; Figure 2 For the present invention Figure 1 A schematic diagram of a flow chart of an embodiment of step S101; Figure 3 For the present invention Figure 1 Schematic diagram of the process of the first embodiment of step S102; Figure 4 For the present invention Figure 1The flowchart of the second embodiment of step S102 is shown in FIG. Figure 5 For the present invention Figure 1 A schematic diagram of a flow chart of an embodiment of step S103; Figure 6 A schematic diagram of the process flow of the same-track multi-overhead crane scheduling method provided by the present invention; Figure 7 A schematic structural diagram of an embodiment of a dispatching device for multiple overhead traveling vehicles on the same track provided by the present invention; Figure 8 A schematic structural diagram of an embodiment of the device provided by the present invention. DETAILED DESCRIPTION
[0019] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0020] In the description of the embodiments of the present invention, unless otherwise specified, "multiple" means two or more than two. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" may mean: A exists alone, A and B exist at the same time, and B exists alone.
[0021] The descriptions of "first", "second", etc. involved in the embodiments of the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the technical features defined as "first" or "second" may explicitly or implicitly include at least one of the features.
[0022] Reference to an "embodiment" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0023] Before describing the embodiments, the following definitions are given for the relevant terms: The Deep Q Network algorithm, abbreviated as DQN algorithm, is an algorithm that combines deep learning and reinforcement learning. It is used to solve reinforcement learning problems in high-dimensional state spaces. The DQN algorithm uses neural networks to approximate and optimize the Q-value function, that is, to estimate the value of taking an action in a given state.
[0024] A greedy algorithm is an algorithm design method based on a greedy strategy, which aims to solve the problem step by step by selecting the option that currently appears to be the best at each step, without considering the overall optimal solution.
[0025] Double DQN (Double Deep Q-Network) is a reinforcement learning algorithm. It is an extended version of DQN and is designed to solve the problem of overestimation in DQN's estimation of Q values. Double DQN solves this problem by using two neural networks, one for selecting the optimal action and the other for estimating the Q value of the action, thereby reducing the impact of overestimation. The two neural networks have exactly the same structure, but different weight parameters.
[0026] The present invention provides a method, device, equipment and storage medium for dispatching multiple overhead travelling vehicles on the same track, which are respectively described below.
[0027] Figure 1 A schematic flow chart of an embodiment of the method for dispatching multiple overhead travelling vehicles on the same track provided by the present invention is shown in FIG. Figure 1 As shown in FIG. 1 , the scheduling method of multiple overhead travelling vehicles on the same track includes: S101, determining the priority of goods to be distributed; S102, generating an initial population according to the priority of the goods to be allocated, and determining the fitness value of each chromosome in the initial population according to the initial population, wherein the fitness value is determined by three factors: the total completion time, the total running distance, and the carbon emission of each overhead crane in the chromosome, and determining the parent chromosome based on the fitness value of the chromosome; S103: Generate a new population according to the parent chromosome, iteratively update the new population to generate a final population, and generate an optimal scheduling plan for the multi-SkyTrain on the same track according to the fitness values of the chromosomes in the final population.
[0028] It should be noted that each overhead crane has completed transport of goods, allocated goods waiting for transportation, and goods to be allocated that are not allocated to an overhead crane. The embodiment selects the best allocation plan for the goods to be allocated, that is, the sequence number and schedule allocated to the overhead crane.
[0029] It should be further explained that sensors are installed on the overhead crane and the warehouse to collect information, and the data is transmitted to the model through communication equipment for data processing to generate the optimal scheduling plan, which is then transmitted to each overhead crane through the communication equipment for operation.
[0030] Compared with the prior art, this embodiment determines the priority of the goods to be allocated, and the priority of the goods to be allocated can solve the overhead crane conflict problem; based on the genetic algorithm, an initial population is generated according to the priority of the goods to be allocated, and the fitness value of the chromosome of the initial population is calculated according to the total completion time, total running distance and carbon emissions of the overhead crane, and the parent chromosome is determined; the parent chromosome is iteratively updated to generate a final population, and according to the fitness value of the chromosome, the chromosome with the smallest fitness value is selected from the final population as the optimal scheduling solution. This embodiment takes into account the real-time environmental status and historical task data, determines the priority of the goods to be allocated, and performs fitness screening on the solution according to the three objectives of the overhead crane completion time, the overhead crane running distance and the overhead crane carbon emissions, and simultaneously performs cross-mutation and iterative updating, thereby realizing multi-objective and multi-level optimization screening of the scheduling solution and obtaining the optimal scheduling solution.
[0031] This embodiment realizes the interaction between decision-making and the environment through reinforcement learning, determines the priority of goods according to different environmental conditions and the type and quality of goods, introduces carbon emission factors in the process of determining the scheduling plan, and combines the total completion time and total running distance of the overhead crane. It not only considers the efficiency of task completion and resource consumption, but also the impact on the environment, and achieves the goals of high efficiency, low loss and environmental protection to optimize the screening of scheduling plans.
[0032] In some embodiments of the present invention, Figure 2 As shown, Figure 2 The present invention provides Figure 1 In the flowchart of step S101 of an embodiment, the real-time environment data includes the current position of the overhead crane, the assigned tasks to be completed, and the material requirements; the priority of the goods to be assigned is determined according to the historical task data, the real-time environment data, the type of goods, and the quality of the goods, including: S201, obtaining historical task data, the current position of the overhead crane, the assigned tasks to be completed, and material requirements; S202, constructing an initial model based on a deep Q network algorithm; S203, determining the current environmental state according to the current position of the overhead crane, the assigned tasks to be completed, and the material requirements; S204, taking the type and quality of the goods to be allocated as input data, training an initial model according to the historical task data, the current environment state and the input data, generating a priority value model, and determining the priority value of the goods to be allocated; S205. Based on a greedy algorithm, the priority value model is trained according to the priority value of the goods to be allocated, a priority division model is generated, and the priority of the goods to be allocated is determined.
[0033] It should be noted that the historical task data includes historical scheduling plan data and the reward returns of historical scheduling plans. The historical task data is used as a basis for reinforcement learning to evaluate the priority value of the goods to be allocated.
[0034] Specifically, firstly, the deep Q network algorithm in the deep reinforcement learning algorithm is used to establish a neural network model, which characterizes the current environmental state by the current position of each overhead crane, the tasks to be completed assigned to the overhead crane, and the demand for each type of product; the neural network is divided into three layers, the input layer, the hidden layer and the output layer. The number of neurons in the hidden layer and the input layer is determined according to the specific warehouse factory situation. The input layer is used to input the current environmental state and the type and quality of goods to be graded, and the output layer outputs the value of each cargo corresponding to different priorities. For example, if there are 3 overhead cranes and 10 types of goods, the input neurons are 29, of which the environmental state is 3*3=9, the types of goods are 10, the quality of goods is 10, the hidden layer is 100, and the activation function is the relu function. After inputting the quantity and quality of the current cargo type and the current environmental state, the neural network returns the value generated by assigning various priorities to the cargo; then, according to the idea of the greedy algorithm, the priority with the largest value is selected as the current cargo priority with a 90% probability, and the priority is randomly selected as the current cargo priority with a 10% probability.
[0035] It should be noted that the Double DQN model is used during training, which uses two neural networks, one neural network as the target network for training, and the other as the computing network for calculating the temporal difference error. After a certain period of time, the parameters of the target network are updated to the computing network.
[0036] This embodiment constructs a priority division model through a deep Q network algorithm and a greedy algorithm, realizes the interaction between decision-making and the environment, determines the priority value of goods for different environmental states and the type and quality of goods, and introduces a neural network to describe the environmental state when the environmental state is limited. Environmental characteristic factors related to priority assessment are selected as inputs to the neural network, such as the demand for various goods, the location of the overhead crane, and the execution of tasks on the overhead crane. The use of deep reinforcement learning methods can well realize the dynamic excellent level determination of various tasks. At the same time, the introduction of the greedy algorithm solves the local optimal choice, thereby finding the global optimal solution.
[0037] In some embodiments of the present invention, Figure 3 As shown, Figure 3 The present invention provides Figure 1 The flowchart of the embodiment of the part of step S102 in the above embodiment, generating an initial population according to the priority of the goods to be allocated, includes: S301, based on the priority division model, construct a scheduling model, based on the genetic algorithm, according to the priority of the goods to be allocated and the current environmental status, assign a gene to each cargo, the value of the gene is set to the crane number, and multiple genes are generated; S302, generating multiple scheduling schemes according to the multiple genes, each scheduling scheme constituting a chromosome; S303. Multiple chromosomes constitute an initial population.
[0038] Specifically, first, a certain number of chromosomes are generated according to the quantity of goods. A chromosome is a scheduling plan, representing a scheduling order or timetable, that is, the order in which each cargo is processed. These chromosomes meet all constraints, such as the maximum load of the overhead crane, the time window of the task, and conflict avoidance.
[0039] In some embodiments of the present invention, Figure 4 As shown, Figure 4 The present invention provides Figure 1 The flowchart of the embodiment of the partial content of step S102 in the above embodiment, according to the initial population, determines the fitness value of each chromosome in the initial population, including: S401. Calculate the weighted sum of each overhead crane in the chromosome according to the total completion time of each overhead crane, the total running distance of the overhead crane, and the carbon emissions of the overhead crane. The weighted sum calculation formula of the overhead crane is: , in, is the weighted sum of the i-th overhead travelling crane, is the total completion time of the i-th overhead crane to complete all tasks, is the total running distance of the ith crane to complete all tasks, The carbon emissions of the ith overhead crane completing all tasks, is the weight factor between distance and completion time, is the weight factor of carbon emissions; S402, according to the weighted sum of the overhead crane, establish the fitness function of the chromosome, the fitness function of the chromosome is: , in, is the fitness value of the chromosome, n is the total number of cranes in the chromosome, is the weighted sum of the nth overhead travelling crane; S403: According to the fitness function of the chromosome, the maximum weighted sum of each overhead travelling vehicle is selected as the fitness value of the chromosome.
[0040] It should be noted that a chromosome represents a scheduling plan, and each chromosome contains multiple overhead cranes. All tasks of each overhead crane include those currently being transported, those that have been assigned to be completed, and those that are about to be assigned; the total completion time of the overhead crane is the cumulative time between the current position of the overhead crane and the destination, the time the overhead crane needs to travel to complete all the assigned tasks to be completed, and the time it needs to travel to complete the tasks that are about to be assigned to the overhead crane; the total operating distance of the overhead crane is the sum of the distances traveled during the time corresponding to the completion of the tasks currently being transported, the tasks that have been assigned to be completed, and the tasks that are about to be assigned; the carbon emissions of the overhead crane are the sum of the carbon emissions during the time corresponding to the completion of the tasks currently being transported, the tasks that have been assigned to be completed, and the tasks that are about to be assigned; based on multiple iterative learning and training, the weight factor between the distance and the completion time and the weight factor of the carbon emissions are determined, and the weighted sum of the overhead crane is calculated. The weighted sum of the overhead crane focuses on the total completion time of the overhead crane, and the weighted sum of the chromosome takes the maximum value of the total completion time of all the overhead cranes in the chromosome, thereby ensuring that within this time, all the overhead cranes of the chromosome have completed the scheduling tasks.
[0041] This embodiment constructs a fitness function based on the total completion time of each overhead crane in the chromosome, the overhead crane running distance and the overhead crane carbon emissions as performance indicators to evaluate the performance of each chromosome. It not only takes into account the efficiency of task completion and resource consumption, but also takes into account the impact of the environment, and achieves the goal of high efficiency, low loss and environmental protection to optimize the screening of scheduling plans.
[0042] In some embodiments of the present invention, determining the parent chromosome based on the fitness value includes: screening out chromosomes whose fitness values are less than a preset fitness threshold and determining them as the parent chromosomes.
[0043] It should be noted that the chromosomes with fitness values less than the preset fitness threshold are screened out. The scheduling scheme performance indicators represented by such chromosomes are better, with good efficiency, low loss and environmental protection. Reproducing these chromosomes as parents helps to pass on excellent characteristics to the next generation, thereby accelerating the convergence to the optimal solution.
[0044] This embodiment selects scheduling schemes with better performance as chromosome parents for reproduction, which helps to pass on excellent characteristics to the next generation, thereby accelerating the speed of convergence to the optimal solution.
[0045] In some embodiments of the present invention, Figure 5 As shown, Figure 5 The present invention provides Figure 1 The flowchart of step S103 of an embodiment includes: S501, performing genetic crossover and mutation operations on the parent chromosomes to generate a new population; S502, using the new population as training data to iteratively train and update the scheduling model, recalculate the fitness value of each chromosome in the new population, determine the parent chromosome of the new population, perform crossover and mutation operations on the parent of the new population, until a preset termination condition is reached, and generate a final population; S503. Select the chromosome with the smallest fitness value from the chromosomes of the final population and generate an optimal scheduling solution.
[0046] Specifically, single-point crossover, multi-point crossover and uniform crossover are used to reproduce the parent chromosomes; single-point crossover is to randomly select a crossover point in the parent chromosome and exchange the parts of the two parent chromosomes after or before the crossover point; multi-point crossover is to randomly select multiple crossover points in the parent chromosome and exchange the chromosome fragments between these crossover points; uniform crossover is to randomly decide for each gene position in the parent chromosome whether to inherit from the first parent chromosome or the second parent chromosome.
[0047] Furthermore, the mutation operation increases the diversity of the population by randomly changing certain genes in the chromosomes to prevent the algorithm from converging to the local optimal solution too early. In the overhead crane scheduling problem, the mutation operation can involve changing the position of one or more tasks in the task list, or in some cases changing the attributes of the tasks.
[0048] Furthermore, by combining crossover and mutation operations, new chromosomes can be generated from the parent chromosomes, and these newly generated chromosomes together with the parent chromosomes constitute a new population. The next round of selection, crossover and mutation operations are performed in the iterative process. The scheduling scheme model constructed by the genetic algorithm is continuously iterated and selected until the preset termination conditions are met, such as reaching the maximum number of iterations, the fitness value reaches the preset optimal threshold, or there is no obvious improvement in the fitness value of several consecutive generations, etc., the iterative selection is terminated to generate the final population.
[0049] This embodiment generates new chromosomes from parent chromosomes through crossover and mutation operations to form a new population, and iteratively trains the scheduling model based on the new population, ultimately generating a chromosome population with good efficiency, low loss and environmental protection.
[0050] In some embodiments of the present invention, after generating the optimal scheduling solution, the method further includes: The fitness value of the optimal scheduling solution is calculated inversely to generate the reward return of the optimal scheduling solution; Based on the time-domain difference method, the parameters of the scheduling model are updated according to the reward returns; Update the overhead crane's dispatch information according to the optimal dispatch plan, and use the optimal dispatch plan as historical task data to update the database.
[0051] Specifically, the fitness value of the optimal scheduling plan is reciprocated, and this reciprocal is used as the reward return of the optimal scheduling plan to give feedback, reflecting the effect of the scheduling plan under the current environmental state. By storing historical reward returns and the optimal scheduling plan, when training the model, the parameters of the scheduling plan model are updated according to the historical reward returns and the current reward returns through the temporal difference method. At the same time, these experiences are randomly adopted to break the correlation between data and improve learning efficiency.
[0052] This embodiment timely reflects the effect of the optimal scheduling in the current environment by timely feeding back the reward return of the optimal scheduling plan, and updates the parameters of the scheduling plan model according to the reward return, thereby improving the learning efficiency and accuracy of the model.
[0053] In order to better illustrate the method of scheduling multiple overhead travelling vehicles on the same track, Figure 6 As shown, Figure 6 A schematic flow chart of a method for dispatching multiple overhead travelling vehicles on the same track according to the present invention is provided.
[0054] In order to better implement the same track multi-overhead train scheduling method in the embodiment of the present invention, based on the same track multi-overhead train scheduling method, correspondingly, Figure 7 As shown, the embodiment of the present invention further provides a dispatching device for multiple overhead traveling vehicles on the same track, and the dispatching device 700 for multiple overhead traveling vehicles on the same track includes: A priority determination module 701 is used to determine the priority of goods to be allocated; The parent chromosome determination module 702 is used to generate an initial population according to the priority of the goods to be allocated, and determine the fitness value of each chromosome in the initial population according to the initial population, wherein the fitness value is determined by three factors: the total completion time, the total running distance and the carbon emission of each overhead crane in the chromosome, and the parent chromosome is determined based on the fitness value of the chromosome; The optimal scheduling solution generation module 703 is used to generate a new population according to the parent chromosome, iteratively update the new population to generate a final population, and generate an optimal scheduling solution for the multi-SkyTrain on the same track according to the fitness values of the chromosomes in the final population.
[0055] The same-track multi-overhead crane dispatching device 700 provided in the above embodiment can implement the technical solution described in the above embodiment of the same-track multi-overhead crane dispatching method. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above embodiment of the same-track multi-overhead crane dispatching method, which will not be repeated here.
[0056] In the embodiment of the present invention, the device for scheduling multiple overhead trains on the same track may be an independent server, or a server network or server cluster composed of servers. For example, the device for scheduling multiple overhead trains on the same track described in the embodiment of the present invention includes but is not limited to a computer, a network host, a single network server, a plurality of network server sets or a cloud server composed of a plurality of servers. The cloud server is composed of a large number of computers or network servers based on cloud computing.
[0057] The present invention also provides a dispatching device for multiple overhead travelling vehicles on the same track. Figure 8 As shown, Figure 8 The structural block diagram of an embodiment of the same track multi-overhead vehicle dispatching device provided by the present invention. The same track multi-overhead vehicle dispatching device 800 can be a computing device such as a mobile terminal, a desktop computer, a notebook, a palm computer, and a server. The same track multi-overhead vehicle dispatching device 800 includes a processor 801 and a memory 802, wherein the memory 802 stores a same track multi-overhead vehicle dispatching program 803.
[0058] In some embodiments, the memory 802 may be an internal storage unit of a computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory 802 may also be an external storage device of the computer device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device. Further, the memory 802 may also include both an internal storage unit of the computer device and an external storage device. The memory 802 is used to store application software and various types of data installed on the computer device, such as program codes for installing the computer device. The memory 802 may also be used to temporarily store data that has been output or is to be output. In one embodiment, the same-track multi-sky train scheduling program 803 may be executed by the processor 801, thereby realizing the same-track multi-sky train scheduling method, device, equipment and storage medium of each embodiment of the present invention.
[0059] In some embodiments, the processor 801 may be a central processing unit (CPU), a microprocessor or other data processing chip, and is used to run program codes or process data stored in the memory 802, such as executing a scheduling program for multiple overhead travelling vehicles on the same track.
[0060] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0061] The above is a detailed introduction to the same-track multi-overhead crane scheduling method, device and image processing equipment provided by the present invention. Specific examples are used in this article to illustrate the principle and implementation mode of the present invention. The description of the above embodiments is only used to help understand the method and core idea of the present invention. At the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation mode and application scope. In summary, the content of this specification should not be understood as limiting the present invention.
Claims
1. A method for dispatching multiple overhead travelling vehicles on the same track, characterized in that: include: Determine the priority of goods to be distributed; Generate an initial population according to the priority of the goods to be allocated, determine the fitness value of each chromosome in the initial population according to the initial population, the fitness value is determined by three factors: the total completion time, the total running distance and the carbon emission of each overhead crane in the chromosome, and determine the parent chromosome based on the fitness value of the chromosome; A new population is generated according to the parent chromosome, the new population is iteratively updated to generate a final population, and an optimal scheduling scheme for multi-SkyTrains on the same track is generated according to the fitness values of the chromosomes in the final population.
2. The method for dispatching multiple overhead travelling vehicles on the same track according to claim 1, characterized in that: Prioritize goods to be distributed, including: Obtain historical task data, current location of the overhead crane, assigned tasks to be completed, and material requirements; Based on the deep Q network algorithm, build the initial model; Determine the current environmental status based on the current position of the overhead crane, the assigned tasks to be completed, and the material requirements; Taking the type of goods to be allocated and the quality of goods to be allocated as input data, training an initial model according to the historical task data, the current environment state and the input data, generating a priority value model, and determining the priority value of the goods to be allocated according to the priority value model; Based on the greedy algorithm, the priority value model is trained according to the priority value of the goods to be allocated, and a priority division model is generated to determine the priority of the goods to be allocated.
3. The method for dispatching multiple overhead travelling vehicles on the same track according to claim 2, characterized in that: According to the priority of the goods to be allocated, an initial population is generated, including: Based on the priority division model, a scheduling model is constructed. Based on the genetic algorithm, a gene is assigned to each cargo according to the priority of the cargo to be assigned and the current environmental status. The value of the gene is set as the crane number, and multiple genes are generated. Generate multiple scheduling schemes based on the multiple genes, each scheduling scheme constituting a chromosome; The plurality of chromosomes constitute an initial population.
4. The method for dispatching multiple overhead travelling vehicles on the same track according to claim 3, characterized in that: Determining the fitness value of each chromosome in the initial population according to the initial population includes: The weighted sum of each crane in the scheduling plan is calculated based on the total completion time of each crane, the total running distance of the crane, and the carbon emissions of the crane. The weighted sum calculation formula of the crane is: , in, is the weighted sum of the i-th overhead travelling crane, is the total completion time of the i-th overhead crane to complete all tasks, is the total running distance of the ith crane to complete all tasks, The carbon emissions of the ith crane completing all tasks, is the weight factor between distance and completion time, is the weight factor of carbon emissions; According to the weighted sum of the overhead crane, the fitness function of the chromosome is established, and the fitness function of the chromosome is: , in, is the fitness value of the chromosome, n is the total number of cranes in the chromosome, is the weighted sum of the nth overhead travelling crane; According to the fitness function of the chromosome, the maximum weighted sum of each overhead travelling vehicle is selected as the fitness value of the chromosome.
5. The method for dispatching multiple overhead travelling vehicles on the same track according to claim 4, characterized in that: Based on the fitness value, the parent chromosome is determined, including: The chromosomes whose fitness values are less than the preset fitness threshold are screened out and determined as parent chromosomes.
6. The method for dispatching multiple overhead travelling vehicles on the same track according to claim 5, characterized in that: According to the parent chromosome, a new population is generated, the new population is iteratively updated to generate a final population, and according to the fitness value of the chromosome in the final population, an optimal scheduling scheme for multiple sky trains on the same track is generated, including: Perform genetic crossover and mutation operations on the parent chromosome to generate a new population; Iteratively training and updating the scheduling scheme model using the new population as training data, recalculating the fitness value of each chromosome in the new population, determining the parent chromosome of the new population, performing crossover and mutation operations on the parent of the new population until a preset termination condition is reached to generate a final population; The chromosome with the smallest fitness value is selected from the chromosomes of the final population to generate an optimal scheduling solution.
7. The method for dispatching multiple overhead travelling vehicles on the same track according to claim 6, characterized in that: After generating the optimal scheduling solution, it also includes: The fitness value of the optimal scheduling solution is calculated to be inverse, and a reward return of the optimal scheduling solution is generated; Based on the time-domain difference method, updating the parameters of the scheduling model according to the reward return; The scheduling information of the overhead crane is updated according to the optimal scheduling scheme, and the optimal scheduling scheme is used as historical task data to update the database.
8. A dispatching device for multiple overhead travelling vehicles on the same track, characterized in that: include: A priority determination module, used to determine the priority of goods to be allocated; A parent chromosome determination module is used to generate an initial population according to the priority of the goods to be allocated, and determine the fitness value of each chromosome in the initial population according to the initial population, wherein the fitness value is determined by three factors: the total completion time, the total running distance, and the carbon emission of each overhead crane in the chromosome, and the parent chromosome is determined based on the fitness value of the chromosome; The optimal scheduling scheme generating module is used to generate a new population according to the parent chromosome, iteratively update the new population to generate a final population, and generate an optimal scheduling scheme for the multi-sky vehicles on the same track according to the fitness values of the chromosomes in the final population.
9. A dispatching device for multiple overhead travelling vehicles on the same track, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the steps of the method for scheduling multiple overhead travelling vehicles on the same track as claimed in any one of claims 1 to 7 are implemented.
10. A storage medium, characterized in that: The storage medium stores computer program instructions, and when the computer program instructions are executed by a computer, the computer is enabled to execute the method for scheduling multiple overhead travelling vehicles on the same track according to any one of claims 1 to 7.
