Task planning optimization method and device for emergency material supply and transportation guarantee

Abstract

Embodiments of the present disclosure provide a task planning optimization method and device for emergency material supply and transportation guarantee, which comprises: constructing an emergency material supply model with the minimum total transportation cost and each task being completed within the task time limit as the objective function; solving the optimal solution of the emergency material supply model and outputting a resource allocation scheme; constructing an information-task association model, associating task decision information with task stages based on historical task information, and outputting a multi-label classification result; taking the resource allocation scheme as input, taking the multi-label classification result as the decision vector of different task stages, taking the task overhead limit and the feasible path without loop in the delivery process as the constraint condition, and taking the minimization of task risk, task time consumption and the number of vehicles as the objective function, constructing a multi-objective optimization model; solving the multi-objective optimization model based on a multi-objective genetic algorithm, and adjusting the model output result according to the real-time feedback of the environment and task parameters to obtain an optimal delivery strategy.

Description

TECHNICAL FIELD

[0001] Embodiments of the present disclosure relate to the technical field of task planning, and in particular, to a task planning optimization method and device for emergency material supply and transportation guarantee, and a computer readable storage medium storing a computer program. BACKGROUND

[0002] Emergency logistics is a special logistics activity for emergency response, which is to meet the needs of materials, personnel and funds for emergency guarantee, and has the characteristics of uncertainty, urgency and irregularity, and its logistics efficiency is usually realized through logistics efficiency. Its content mainly includes the storage and management of emergency materials, the transportation and distribution of emergency materials, and the construction of emergency logistics organization mechanism.

[0003] In the modern emergency guarantee system, due to the continuous emergence and intervention of unmanned and intelligent devices, the emergency guarantee task presents a highly complex and changeable situation, with long task time span, wide action region, and many involved guarantee elements, which puts forward new requirements for the rapid response capability of rescue forces. How to accurately plan the task, schedule the resources and perform the monitoring in a highly complex and dynamic environment still faces great challenges.

[0004] At present, the complexity of emergency guarantee tasks not only comes from the diversity of the tasks themselves, but also comes from the influence of various sudden factors on the tasks, which needs to consider the urgency of the task, the availability of resources, the priority of materials, the safety during transportation and other factors. These factors often restrict each other, so a more complex multi-objective optimization method is needed for comprehensive trade-off and decision-making. The Chinese patent document with publication number CN113705970A discloses an HTN planning method for emergency logistics resource task matching. In the case where the demand, transportation time and unit material transportation cost are all uncertain, a multi-objective model is established to minimize the emergency time, cost and material safety transportation period. The theory of interval numbers and triangular fuzzy numbers is used to solve the multi-objective model. This scheme uses fuzzy mathematical theory to deal with uncertainty, but it relies on a pre-set hierarchical task structure and a fixed decision path, which is difficult to adjust the decision and task path in real time when facing complex dynamic environment and uncertain factors. SUMMARY

[0005] The embodiments described herein provide a task planning optimization method and device for emergency material supply and transportation guarantee, and a computer readable storage medium storing a computer program.

[0006] According to a first aspect of the present disclosure, a task planning optimization method for emergency material supply and transportation guarantee is provided, which comprises: constructing an emergency material supply model with a minimum total transportation cost and each task being completed within a task time limit as an objective function; solving the optimal solution of the emergency material supply model to output a resource allocation scheme; constructing an information-task association model to associate task decision information with a task phase based on historical task information, and output a multi-label classification result; taking the resource allocation scheme as an input, taking the multi-label classification result as a decision vector of different task phases, taking the task overhead limit and the feasible path without loop in the delivery process as constraint conditions, and taking the minimum task risk, the minimum task time consumption and the minimum number of vehicles as objective functions, constructing a multi-objective optimization model of emergency logistics tasks; and solving the multi-objective optimization model based on a multi-objective genetic algorithm, and adjusting the model output result according to the real-time feedback of environmental parameters and task parameters to obtain an optimal delivery strategy in a real-time environment.

[0007] In some embodiments of the present disclosure, the objective function of the emergency material supply model is constructed as:

[0008]

[0009] s.t.P={p j |T ij ≤T j (j=1,2,...,n)}

[0010] In the formula, Z is the total transportation cost, m is the number of material supply points, n is the number of tasks, C ij is the unit transportation price, X ij is the transportation quantity from the material supply point i to the guarantee object j, T j is the guarantee time limit of each task, T ij is the transportation time from the material supply point i to the guarantee object j, p j is a single task, P is a task set; and

[0011] The constraint conditions of the emergency material supply model are determined as:

[0012]

[0013] In the formula, a i is the quantity of materials stored at each material supply point i, b j is the material demand of the guarantee object j in each task, and X ij is the transportation quantity from the material supply point i to the guarantee object j.

[0014] In some embodiments of the present disclosure, a unit freight rate and a transportation time are calculated for each transportation path, a transportation path with the lowest unit freight rate and the shortest transportation time is selected as an initial configuration scheme for allocation of the materials, until all the materials and demands are satisfied, the initial configuration scheme is completed; the initial configuration scheme is optimally discriminated according to the inspection numbers of all non-basic variables, if all the inspection numbers are not less than 0, the current scheme is an optimal solution; and if there is an inspection number less than 0, the current scheme is not an optimal solution, the scheme is adjusted and the optimal discrimination is continuously performed until all the inspection numbers are greater than 0, an optimal solution of the emergency material supply model objective function is obtained, and a resource configuration scheme is output.

[0015] In some embodiments of the present disclosure, an optimal solution of the emergency material supply model objective function is obtained by the following steps: step 1, selecting a variable corresponding to the minimum negative inspection number as a basic variable, comparing the ratios of the variables in each constraint condition, and selecting a variable with the smallest ratio as a non-basic variable; step 2, increasing the value of the basic variable and adjusting the non-basic variable to 0, and recalculating the inspection numbers of each variable; step 3, repeating steps 1-2 until all the inspection numbers are greater than 0, and an optimal solution of the emergency material supply model objective function is obtained.

[0016] In some embodiments of the present disclosure, a data set containing task decision information and task stages is constructed based on historical task information, the data set is used as training data of a multi-label learning algorithm, each data sample contains a feature vector of the task decision information and a label vector of the task stage; any one of a k-nearest neighbor-based multi-label classification algorithm, a kernel-based multi-label classification algorithm, and a deep semantic matching multi-label classification algorithm is used to construct an information-task association model, the information-task association model is trained based on the training data, and a trained information-task association model is obtained; and a real-time task information feature is input into the trained information-task association model, and a multi-label classification result is output.

[0017] In some embodiments of the present disclosure, an expression of a multi-objective optimization model is constructed as follows:

[0018] min y = [f1(x), f2(x), f3(x),..., f M (x)]

[0019] In the formula, x = (x1, x2..., x n ) ∈ D is a decision vector formed by a multi-label learning classification result, D is a decision space formed by the decision vector, y = (f1(x), f2{x),..., f M ) ∈ Y represents a target space formed by an objective function of the multi-objective optimization model, the objective function of the multi-objective optimization model includes minimizing a task risk, minimizing a task time consumption, and minimizing a number of vehicles; and

[0020] The task overhead limit constraint of the multi-objective optimization model is determined:

[0021]

[0022] wherein, C all is the sum of the costs of all routes in the task road network; C k is the cost of a single route; is the driving cost of the kth route, is the cost per unit time, is the time of delay; is the cost caused by the safety factor; toll k is the additional cost caused by unpredictable factors.

[0023] In some embodiments of the present disclosure, an initial population of a genetic algorithm is generated according to a multi-label classification result output by an information-task association model, environmental information, and echelon position information; an adaptive value of each individual in the initial population is calculated based on a target function of a multi-objective optimization model by a Pareto optimization method; the individuals in the population are selected, crossed, and mutated according to the adaptive value of each individual; the decision variables and the target function of the multi-objective optimization model are adjusted according to real-time environmental parameters and task parameters; and the Pareto front is updated according to the adjusted decision variables and the target function, and the solution of the multi-objective optimization model is iteratively optimized until a maximum number of iterations is reached, and an optimal delivery strategy under real-time environment is output.

[0024] In some embodiments of the present disclosure, each individual in the initial population contains task allocation, path selection, and selection of a transportation tool, the environmental parameters include weather conditions, traffic conditions, resource availability, and emergencies, and the task parameters include the priority and time limit of the task.

[0025] According to a second aspect of the present disclosure, a task planning optimization device for emergency material supply and transportation support is provided, the device comprising at least one processor and at least one memory storing a computer program. When the computer program is executed by the at least one processor, the device: constructs an emergency material supply model with the minimum total freight and the completion of each task within the task time limit as the objective function; solves the emergency material supply model for the optimal solution and outputs a resource allocation plan; constructs an information-task association model, associates task decision information with task stages based on historical task information, and outputs a multi-label classification result; uses the resource allocation plan as input, uses the multi-label classification result as the decision vector for different task stages, uses task cost limit and loop-free feasible path during delivery as constraints, and constructs a multi-objective optimization model for emergency logistics tasks with the objective function of minimizing task risk, minimizing task time, and minimizing the number of carriers; and solves the multi-objective optimization model based on a multi-objective genetic algorithm, and adjusts the model output according to real-time feedback of environmental parameters and task parameters to obtain the optimal delivery strategy in a real-time environment.

[0026] In some embodiments of the present disclosure, when the computer program is executed by at least one processor, the computer program causes the apparatus to construct an emergency material supply model by performing the following operations: constructing an objective function of the emergency material supply model:

[0027]

[0028] stP={p j |T ij ≤T j (j=1, 2, ..., n)}

[0029] In the formula, Z is the total freight, m ​​is the number of material supply points, n is the number of tasks, C ij is the unit freight rate, X ij is the transportation volume from material supply point i to support object j, T j The guarantee time limit for each mission, T ij is the transportation time from material supply point i to support object j, p j is a single task, P is a set of tasks; and

[0030] Determine the constraints of the emergency supplies supply model:

[0031]

[0032] Where a i The amount of supplies stored at each supply point i, b j is the material demand of object j for each mission, X ij is the transportation volume from material supply point i to support object j.

[0033] In some embodiments of the present disclosure, the computer program, when executed by the at least one processor, causes the apparatus to output the resource allocation scheme by: calculating a unit freight and a transportation time for each transportation path, selecting a transportation path with the lowest unit freight and the shortest transportation time as an initial configuration scheme for allocation of the materials, until all the materials and demands are satisfied, completing the initial configuration scheme; performing optimal discrimination on the initial configuration scheme according to the inspection numbers of all non-basic variables, if all the inspection numbers are not less than 0, the current scheme is an optimal solution; and if there is an inspection number less than 0, the current scheme is not an optimal solution, then adjusting the scheme and continuing to perform optimal discrimination until all the inspection numbers are greater than 0, obtaining an optimal solution of the emergency material supply model objective function, and outputting the resource allocation scheme.

[0034] In some embodiments of the present disclosure, the computer program, when executed by the at least one processor, causes the apparatus to obtain an optimal solution of the emergency material supply model objective function by: step 1, selecting a variable corresponding to the smallest negative inspection number as a basic variable, comparing the ratios of the variables in each constraint condition, and selecting the variable with the smallest ratio as a non-basic variable; step 2, increasing the value of the basic variable, adjusting the non-basic variable to 0, and recalculating the inspection numbers of each variable; and repeating steps 1-2 if there is still an inspection number less than 0 until all the inspection numbers are greater than 0, obtaining an optimal solution of the emergency material supply model objective function.

[0035] In some embodiments of the present disclosure, the computer program, when executed by the at least one processor, causes the apparatus to associate the task decision information with the task stage by: constructing a data set containing the task decision information and the task stage based on historical task information, using the data set as training data of a multi-label learning algorithm, each data sample containing a feature vector of the task decision information and a label vector of the task stage; using any one of a k-nearest neighbor-based multi-label classification algorithm, a kernel-based multi-label classification algorithm, and a deep semantic matching multi-label classification algorithm to construct an information-task association model, training the information-task association model based on the training data to obtain a trained information-task association model; and inputting real-time task information features into the trained information-task association model to output a multi-label classification result.

[0036] In some embodiments of the present disclosure, the computer program, when executed by the at least one processor, causes the apparatus to construct a multi-objective optimization model of an emergency logistics task by: constructing an expression of the multi-objective optimization model:

[0037] miny=[f1(x),f2(x),f3(x),...,f M (x)]

[0038] wherein x = (x1, x2,..., x n ) e D is a decision vector formed by multi-label learning classification results, D is a decision space formed by decision vectors, y = (f1(x), f2(x),..., f M ) e Y represents a target space formed by objective functions of a multi-objective optimization model, the objective functions of the multi-objective optimization model include minimizing task risk, minimizing task time consumption, and minimizing the number of vehicles; and

[0039] determining a task overhead limit constraint condition of the multi-objective optimization model:

[0040]

[0041] wherein C all is the sum of the costs of all routes in the task road network; C k is the cost of a single route; is the driving cost of the kth route, is the cost per unit time, is the time of delay; is the cost caused by the safety factor; toll k is the additional cost caused by unpredictable factors.

[0042] In some embodiments of the present disclosure, the computer program, when executed by at least one processor, causes the apparatus to obtain an optimal delivery strategy under a real-time environment by:

[0043] generating an initial population of a genetic algorithm based on the multi-label classification results output by the information-task association model, the environmental information, and the echelon position information; calculating the fitness value of each individual in the initial population based on the objective functions of the multi-objective optimization model based on the Pareto optimization method; performing selection, crossover, and mutation operations on the individuals in the population according to the fitness value of each individual; adjusting the decision variables and the objective functions of the multi-objective optimization model according to real-time environmental parameters and task parameters; and updating the Pareto front based on the adjusted decision variables and the objective functions, iteratively optimizing the solution of the multi-objective optimization model until the maximum number of iterations is reached, and outputting the optimal delivery strategy under the real-time environment.

[0044] According to a third aspect of the present disclosure, there is provided a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the task planning optimization method for emergency material supply and transportation guarantee according to the first aspect of the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments will be briefly described in the following. It should be known that the drawings described below only relate to some of the embodiments of the present disclosure, rather than limit the present disclosure, in which:

[0046] Figure 1 A flowchart of a task planning optimization method 100 for emergency material supply and transportation guarantee according to an embodiment of the present disclosure is shown;

[0047] Figure 2 A schematic diagram of a learning process of an information-task association model according to an embodiment of the present disclosure is shown;

[0048] Figure 3 A flowchart of a multi-objective genetic algorithm according to an embodiment of the present disclosure is shown;

[0049] Figure 4 A schematic block diagram of a task planning optimization device 400 for emergency material supply and transportation guarantee according to an embodiment of the present disclosure is shown.

[0050] It should be noted that the elements in the drawings are schematic and not drawn to scale. DETAILED DESCRIPTION

[0051] In order to make the purposes, technical solutions and advantages of the embodiments of the present disclosure clearer, the technical solutions of the embodiments of the present disclosure will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by a person of ordinary skill in the art without creative effort also belong to the scope of protection of the present disclosure.

[0052] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this subject matter belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the specification and relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein. Additionally, terms such as "first" and "second" are used merely to distinguish one component (or portion of a component) from another.

[0053] The scheme can effectively generate a comprehensive emergency logistics guarantee scheme through the cooperation of the emergency material supply model and the multi-objective optimization model of emergency logistics tasks, and ensure that resources can be allocated and transportation guarantee can be carried out quickly and efficiently in emergencies. The core goal of the emergency material supply model is to calculate and plan how to effectively allocate emergency resources to meet the material needs of each demand point in disasters or emergencies. The multi-objective optimization model of emergency logistics tasks is responsible for planning the transportation path, scheduling the transportation tools and ensuring the smooth progress of the transportation process.

[0054] Figure 1 A flowchart of a task planning optimization method 100 for emergency material supply and transportation guarantee according to an embodiment of the present disclosure is shown. As shown in Figure 1 frame S110, an emergency material supply model is constructed with the total transportation cost being the minimum and each task being completed within the task time limit as the objective function.

[0055] In an embodiment of the present disclosure, m material supply points are pre-deployed, and the amount of material stored in each material supply point is a i (i = 1, 2,..., m). It is assumed that the scheme of the guarantee action is composed of n tasks, the task set is P = {p n}, the guarantee time limit specified by each task is T j , the material demand of the guarantee object j of each task is b j (j = 1, 2,..., n), the transportation time from the material supply point i to the guarantee object j is T ij , the unit transportation price is c ij , and the transportation amount from the material supply point i to the guarantee object j is X ij .

[0056] In order to minimize the total transportation cost and complete each task within the task time limit, the objective function of the emergency material supply model is constructed as follows:

[0057]

[0058] s.t.P = {p j |T ij ≤T j (j = 1, 2,..., n)}

[0059] wherein Z is the total transportation cost, m is the number of material supply points, n is the number of tasks, C ij is the unit transportation price, X ij is the transportation amount from the material supply point i to the guarantee object j, T j is the guarantee time limit specified by each task, T ij is the transportation time from the material supply point i to the guarantee object j, and p jP is a set of tasks for a single mission.

[0060] The constraint condition of the emergency material supply model is determined as:

[0061]

[0062] In the formula, a i is the quantity of the material stored at each material supply point i, b j is the quantity of the material required by each task guarantee object j, X ij is the quantity of the material transported from the material supply point i to the guarantee object j.

[0063] Then, in block S120, the optimal solution of the emergency material supply model is solved, and the resource allocation scheme is output.

[0064] First, the initial allocation scheme is determined. The unit transportation price and transportation time can be calculated for each transportation path, and the path with the lowest unit transportation price and the shortest transportation time is preferentially selected for material allocation, until all materials and requirements are met, and the initial allocation scheme is completed. That is, according to the minimum unit transportation price and the shortest transportation time of the selected path, the available materials are allocated to the paths with the lowest price and the shortest transportation time, until the requirements or processes of a path are completely met. When the requirements or supply of a path are completely met, the next path with the lowest unit transportation price and the shortest transportation time is selected, and the material allocation is continued until all materials and requirements are met. This method can quickly find a feasible solution without excessive calculation and complex judgment. However, it should be noted that this method may not be able to find a global optimal solution, so after obtaining the initial scheme, the initial scheme needs to be optimally discriminated.

[0065] The test number of the non-basic variable is a key tool for judging whether the scheme is optimal. The initial allocation scheme can be optimally discriminated according to the test numbers of all non-basic variables. If all test numbers are not less than 0, the current scheme is the optimal solution. Specifically, the test number δ ij of each non-basic variable is calculated, that is, a non-basic variable and several basic variables form a closed loop. The sum of the unit transportation prices of the odd vertices of the closed loop minus the sum of the unit transportation prices of the even vertices is the test number of the non-basic variable. If all test numbers δ

[0066] The adjustment scheme is as follows:

[0067] Step 1: Select the variable with the smallest negative test value as the entry variable. Compare the ratios of the variables in each constraint and select the variable with the smallest ratio as the exit variable. Selecting the path with the smallest absolute negative test value indicates that it is the path most likely to reduce costs. This adjustment is done by reallocating certain items from the current path to the new path, in order to achieve a positive test value. This adjustment involves adding or removing some paths to minimize transportation costs across the entire system.

[0068] Step 2: Increase the value of the input variable, adjust the output variable to 0, and recalculate the test number for each variable. After the adjustment, you need to recalculate the remaining supplies and demand for each supply point and task, and recalculate the corresponding unit freight rate. Continue to check whether all test numbers are greater than 0.

[0069] Step 3: If any test values ​​are still less than 0, repeat steps 1-2 until all test values ​​are greater than 0, resulting in the optimal solution for the objective function of the emergency supply model. If any test values ​​are negative, continue adjusting the plan. After multiple rounds of adjustments, if all test values ​​are greater than 0, the final resource allocation plan is the optimal solution, meeting all material needs while minimizing transportation costs and ensuring that each task is completed within the specified time limit.

[0070] Then refer to Figure 1 As shown, in block S130 , an information-task association model is constructed, task decision information is associated with task stages based on historical task information, and a multi-label classification result is output.

[0071] The relationship between task decision information and task stages is modeled using multi-label learning. Different stages of a task may require multiple different decisions, and the decisions at each stage need to be classified based on multiple features. Multi-label learning is very useful in this case because it allows each task stage to have multiple decision labels simultaneously.

[0072] According to one embodiment of the present disclosure, a task in an information-task association model typically consists of multiple stages, each with multiple decisions. The decision-making information for each task stage can be described by multiple features. By learning from historical task information, the different stages of a task can be accurately classified. These classification results provide decision variables for subsequent multi-objective optimization models.

[0073] Figure 2 Schematic diagram of the learning process of the information-task association model according to an embodiment of the present disclosure is shown. Figure 2As shown, the input data of the information-task association model includes basic information of historical tasks (such as task priority, nature of transported goods, time requirement, etc.) and environmental data (such as weather, road conditions, traffic flow, etc.).

[0074] According to one embodiment of the present disclosure, a dataset containing task decision information and task stages is constructed based on historical task information, and the dataset is used as training data of a multi-label learning algorithm, each data sample containing a feature vector of task decision information and a label vector of task stages. For example, the feature vector describes various decision information of the task stage (such as task time limit, resource condition, weather, traffic condition, etc.), and the label vector of the task stage includes path selection, tool selection, resource allocation, etc. Referring to Figure 2 As shown, information feature extraction is performed on the input data to obtain multiple label sets of information and task association. The model uses a multi-label classification learning algorithm, such as any one of decision tree, random forest, k-nearest neighbor-based multi-label classification algorithm, kernel-based multi-label classification algorithm, and deep semantic matching multi-label classification algorithm. The information-task association model is trained based on the training data to obtain a trained information-task association model. Real-time task information features are input into the trained information-task association model, and a multi-label classification result is output, and the task decision space is updated in real time.

[0075] For example, assuming that there is a new task stage with input features [weather: cloudy, traffic: congested, time limit: relaxed, tool availability: medium], the trained multi-label classification model is used for prediction, and the output label vector is:

[0076] Predicted label (Y _ pred): [path B, path A] (path selection)

[0077] Predicted label (Y _ pred): [tool 2, tool 1] (tool selection)

[0078] Predicted label (Y _ pred): [resource 2, resource 1] (resource allocation)

[0079] Subsequently, a multi-objective optimization model is constructed based on the information-task association model. Referring back to Figure 1 As shown, in block S140, a resource allocation scheme is input, the multi-label classification result is used as the decision vector of different task stages, the task overhead limit and the feasible path without loop in the delivery process are used as constraint conditions, and the minimum task risk, the minimum task time consumption, and the minimum number of vehicles are used as objective functions to construct a multi-objective optimization model of the emergency logistics task.

[0080] According to one embodiment of the present disclosure, the multi-objective optimization model can simultaneously optimize multiple conflicting objective functions and constraint conditions. For example, M objective functions include task risk minimization, transportation time consumption minimization, and vehicle quantity minimization, n decision variables include task allocation scheme, transportation path, selection of transportation tools, etc., and K constraint conditions include road network cost constraint, path constraint, capacity constraint, etc. The expression of the multi-objective optimization model for emergency logistics transportation guarantee is:

[0081] min y=f(x)=[f1(x), f2(x), f3(x),..., f M (x)]

[0082] In the formula, x=(x1, x2..., x n )∈D is a decision vector formed by multi-label learning classification results, D is a decision space formed by the decision vector, y=(f1(x), f2(x),..., f M )∈Y represents a target space formed by objective functions of the multi-objective optimization model, and the objective functions of the multi-objective optimization model include minimizing task risk, minimizing task time consumption, and minimizing vehicle quantity. The expression simultaneously satisfies the generalized cost minimization constraint condition of the road network:

[0083]

[0084] Wherein, C all is the sum of the costs of all lines in the task road network; C k is the cost of a single line; is the driving cost of the kth line, is the cost per unit time, is the time of delay; is the cost of the safety factor; toll k is the additional cost caused by unpredictable factors.

[0085] Considering that in emergency logistics, the road network environment usually changes dynamically, such as traffic congestion, natural disasters, etc., which can cause uncertainty of the path, the embodiment of the present disclosure increases the influence of factors such as traffic conditions, weather changes, road damage, etc. on the basis of the traditional path cost model considering time and distance, and evaluates the cost of each path according to the congestion status, traffic capacity, etc. of the road network, so as to cope with emergencies and dynamic changes.

[0086] Returning to Figure 1 , in block S150, the multi-objective optimization model is solved based on the multi-objective genetic algorithm, and the model output result is adjusted according to the real-time feedback of the environmental parameters and the task parameters, to obtain the optimal delivery strategy under the real-time environment.

[0087] Considering the multi-objective, nonlinearity, and high dimensionality of the multi-objective optimization model, traditional optimization algorithms (such as Dijkstra's shortest path algorithm) cannot effectively solve the multi-objective optimization problem in such a dynamic environment. Therefore, the genetic algorithm is used as the optimization solution method, and is further extended to the multi-objective genetic algorithm (Multi-Objective Genetic Algorithm, MOGA) to handle multiple conflicting objectives.

[0088] Figure 3 A flowchart of a multi-objective genetic algorithm according to an embodiment of the present disclosure is shown. Referring to Figure 3 As shown, first, the multi-label classification results output according to the information-task association model, the environment information, and the echelon position information are used to generate the initial population of the genetic algorithm. Each individual in the population represents a possible transportation scheme, including path selection, task allocation, use of transportation tools, etc. The merits of each solution are evaluated by calculating the objective function value, i.e., the fitness value. The fitness value of each solution is determined by multiple objectives such as task risk, transportation time consumption, and number of vehicles. The fitness value of each individual in the initial population can be calculated based on the Pareto optimization method through the objective function of the multi-objective optimization model. That is, the optimal solution set of each objective is directly obtained by the method of Pareto frontier. The merits of the solution are selected by judging the dominance relationship of the solution, without weighting the objectives. During the fitness evaluation process, feedback from real-time environmental parameters (such as weather, traffic, road conditions) and task parameters (such as task priority, time requirements, etc.) needs to be considered.

[0089] According to the fitness value of each individual, the individuals in the population are selected, crossed, and mutated. Specifically, individuals with better fitness are selected to participate in the crossover and mutation operations. The crossover operation is to exchange the gene parts of two parent individuals to generate new individuals (offspring), and the mutation operation is to randomly change the genes of the individuals within a small range to increase the explorability of the solution space and prevent the algorithm from falling into local optimum.

[0090] In each iteration of the genetic algorithm, the environmental conditions may change. The decision variables and objective functions of the multi-objective optimization model can be adjusted according to the real-time environmental parameters and task parameters. The environmental parameters include weather conditions, traffic conditions, resource availability, emergencies, etc., and the task parameters include the priority and time limit of the task, etc. The Pareto frontier is updated according to the adjusted decision variables and objective functions, and the solution of the multi-objective optimization model is iteratively optimized until the maximum number of iterations is reached, and the optimal delivery strategy under real-time environment is output. After multiple iterations and real-time adjustments, the MOGA model outputs a set of Pareto optimal solutions, i.e., the trade-off solutions between multiple objectives.

[0091] That is, before reaching the destination, each change in environmental information triggers a re-planning of the path, i.e., performing the genetic algorithm once. If the environmental change occurs during the execution of the genetic algorithm, the algorithm immediately suspends the current genetic algorithm iteration, clears the relevant variables and redundant data, reads in the new environmental information and team position information, reinitializes and completes the calculation.

[0092] It should be noted that the decision maker can select a scheme that best meets the current task requirements and environmental conditions from these Pareto optimal solutions as the final transportation support scheme according to actual circumstances (such as risk tolerance, time requirements, etc.).

[0093] Through the multi-objective optimization model based on the multi-objective genetic algorithm, real-time path optimization, task allocation and resource scheduling can be achieved in a complex and dynamic emergency logistics environment. Under the constantly changing environment and task parameters, the genetic algorithm can quickly adjust the strategy to ensure timely updating and execution of the optimal delivery scheme, thereby ensuring efficient and accurate completion of emergency tasks.

[0094] Figure 4 is a schematic block diagram of a task planning optimization device 400 for emergency material supply and transportation support according to an embodiment of the present disclosure. Referring to Figure 4 , the device 400 can include a processor 410 and a memory 420 storing a computer program. When the computer program is executed by the processor 410, the device 400 can perform the steps of the method 100 as shown in Figure 1 . In one example, the device 400 can be a computer device or a cloud computing node. The device 400 can construct an emergency material supply model with the objective function of minimizing the total transportation cost and completing each task within the task time limit. The device 400 can solve the optimal solution for the emergency material supply model and output a resource allocation scheme. The device 400 can construct an information-task association model, associate task decision information with task stages based on historical task information, and output a multi-label classification result. The device 400 can construct a multi-objective optimization model for emergency logistics tasks by taking the resource allocation scheme as input, the multi-label classification result as the decision vector for different task stages, the task overhead limit and the feasible path without loops in the delivery process as constraint conditions, and minimizing the task risk, minimizing the task time consumption, and minimizing the number of vehicles as objective functions. The device 400 can solve the multi-objective optimization model based on the multi-objective genetic algorithm and adjust the model output result based on real-time feedback of environmental parameters and task parameters to obtain the optimal delivery strategy under real-time environment.

[0095] In some embodiments of the present disclosure, the device 400 can construct an emergency material supply model by constructing the objective function of the emergency material supply model:

[0096]

[0097] s.t.P={p j |T ij ≤T j (j=1,2,...,n)}

[0098] wherein Z is the total transportation cost, m is the number of supply points, n is the number of tasks, C ij is the unit transportation cost, X ij is the transportation amount from the supply point i to the object j, T j is the time limit of each task, T ij is the transportation time from the supply point i to the object j, p j is a single task, and P is a task set; and

[0099] The constraint conditions of the emergency material supply model are determined as follows:

[0100]

[0101] wherein a i is the amount of materials stored at each supply point i, b j is the material demand of the object j in each task, and X ij is the transportation amount from the supply point i to the object j.

[0102] In some embodiments of the present disclosure, the device 400 can output the resource allocation scheme by: calculating the unit transportation cost and the transportation time for each transportation path, selecting the transportation path with the lowest unit transportation cost and the shortest transportation time as an initial configuration scheme for material allocation, until all materials and demands are satisfied, completing the initial configuration scheme; performing optimal discrimination on the initial configuration scheme according to the inspection numbers of all non-basic variables, if all the inspection numbers are not less than 0, the current scheme is the optimal solution; and if there is an inspection number less than 0, the current scheme is not the optimal solution, then adjusting the scheme and continuing to perform optimal discrimination until all the inspection numbers are greater than 0, obtaining the optimal solution of the objective function of the emergency material supply model, and outputting the resource allocation scheme.

[0103] In some embodiments of the present disclosure, the device 400 can obtain the optimal solution of the emergency material supply model objective function by the following operations: step 1, selecting the variable corresponding to the minimum negative test number as the in-base variable, comparing the ratio of the variable in each constraint condition, and selecting the variable with the smallest ratio as the out-base variable; step 2, increasing the value of the in-base variable, adjusting the out-base variable to 0, and recalculating the test number of each variable; step 3, repeating steps 1-2 if there are still test numbers less than 0 until all test numbers are greater than 0, obtaining the optimal solution of the emergency material supply model objective function.

[0104] In some embodiments of the present disclosure, the device 400 can associate task decision information with a task phase by: constructing a data set containing task decision information and task phases based on historical task information, using the data set as training data for a multi-label learning algorithm, and each data sample containing a feature vector of task decision information and a label vector of task phase; using any one of a k-nearest neighbor-based multi-label classification algorithm, a kernel-based multi-label classification algorithm, and a deep semantic matching multi-label classification algorithm to construct an information-task association model, training the information-task association model based on the training data to obtain a trained information-task association model; and inputting real-time task information features into the trained information-task association model, and outputting a multi-label classification result.

[0105] In some embodiments of the present disclosure, the device 400 can construct a multi-objective optimization model of an emergency logistics task by: constructing an expression of the multi-objective optimization model:

[0106] miny=[f1(x),f2(x),f3(x),...,f M (x)]

[0107] In the formula, x=(x1,x2...,x n )∈D is a decision vector formed by a multi-label learning classification result, D is a decision space formed by the decision vector, y=(f1(x),f2(x),...,f M )∈Y represents a target space formed by the objective function of the multi-objective optimization model, the objective function of the multi-objective optimization model includes minimizing task risk, minimizing task time consumption, and minimizing the number of vehicles; and

[0108] The task overhead limit constraint condition of the multi-objective optimization model is determined as:

[0109]

[0110] In the formula, C all is the sum of the costs of all lines in the task road network; C k is the cost of a single line. is the travel cost of the kth route, is the cost per unit time, is the time of delay; is the cost generated by the safety factor; toll k is the additional cost caused by unpredictable factors.

[0111] In some embodiments of the present disclosure, the device 400 can obtain the optimal delivery strategy in real-time environment by the following operations:

[0112] According to the multi-label classification result output by the information-task association model, the environment information and the echelon position information, an initial population of genetic algorithm is generated; the fitness value of each individual in the initial population is calculated based on the objective function of the multi-objective optimization model by the Pareto optimization method; the individuals in the population are selected, crossed and mutated according to the fitness value of each individual; the decision variables and the objective function of the multi-objective optimization model are adjusted according to the real-time environment parameters and the task parameters; and the Pareto front is updated according to the adjusted decision variables and the objective function, and the solution of the multi-objective optimization model is iteratively optimized until the maximum number of iterations is reached, and the optimal delivery strategy in real-time environment is output.

[0113] In embodiments of the present disclosure, the processor 410 can be, for example, a central processing unit (CPU), a microprocessor, a digital signal processor (DSP), a processor based on a multi-core processor architecture, etc. The memory 420 can be any type of memory implemented using data storage technology, including but not limited to random access memory, read-only memory, semiconductor-based memory, flash memory, disk storage, etc.

[0114] In addition, in embodiments of the present disclosure, the device 400 can also include an input device 430, such as a keyboard, a mouse, etc., for inputting historical task information, etc. In addition, the device 400 can also include an output device 440, such as a display, etc., for outputting resource allocation schemes, multi-label classification results and optimal delivery strategies, etc.

[0115] In other embodiments of the present disclosure, a computer readable storage medium storing a computer program is also provided, wherein the computer program can implement the steps of the task planning optimization method 100 for emergency material supply and transportation guarantee as shown in Figure 1

[0116] ​In summary, the task planning optimization method and device for emergency material supply and transportation guarantee according to the embodiments of the present disclosure can select the optimal transportation path by considering the unit transportation price and transportation time simultaneously, avoid the incomplete optimal decision made under a single target, and more comprehensively improve the transportation efficiency and reduce the cost; by integrating and semantically matching the multi-source emergency logistics task data, the correlation between resource supply and transportation in a complex logistics system can be revealed, and comprehensive data support for emergency guarantee is provided; for the complexity, nonlinearity and emergent problems in emergency logistics guarantee, the multi-objective genetic algorithm is used for dynamic path planning, and the road network generalized cost evaluation technology is combined, so that the configuration of emergency logistics resources and the transportation path can be optimized under the background of multiple targets and multiple constraints, the task efficient execution is ensured, the output of the optimization model is adjusted based on the real-time feedback of the environmental parameters and the task parameters, the decision information can be updated in real time, the changes of the environment and the task demand are adapted, and the overall guarantee capability of the emergency logistics system is ultimately improved.

[0117] The flow and block diagrams in the drawings show the architectural, functional, and operational aspects of possible implementations of apparatuses and methods according to various embodiments of the present disclosure. In this regard, each block in the flow and block diagrams can represent a module, a segment, or a portion of code that comprises one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession can in fact be executed substantially concurrently or the blocks can sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block and / or flow diagrams and combinations of blocks in the block and / or flow diagrams can be implemented by dedicated hardware-based systems that perform the specified functions or acts, or can be implemented by a combination of dedicated hardware and computer instructions.

[0118] Unless specifically stated otherwise, and as can be apparent from the context, use herein and in the appended claims of a singular form of a word that includes one or more words of a pluralopposite also include the plural and vice versa. So, for example, a discussion regarding "a" can include multiple instances of "a" and vice versa. Similarly, the use of the word "comprises" or "comprising" is intended to mean that a stated feature, integer, step, or group of features, integers, or steps is included, but not to the exclusion of the presence or addition of one or more other features, integers, steps, or groups thereof. Also, the terms "comprises", "comprising", "includes", "including" and the like can be used synonymously. Additionally, the term "exemplary" as used herein, particularly in the context of an example, is intended to indicate that a particular implementation has some or all of the attributes, features, or benefits described or in some other way exemplifies the described implementation. Moreover, the use of the term "or" in the context of some embodiments is to be interpreted as an inclusive "or" rather than an exclusive "or" unless otherwise noted. That is, unless specified otherwise, a combination of elements or options is within the scope of some embodiments, whether or not the combination is used or described herein.

[0119] Further aspects and ranges of adaptation become apparent from the description provided herein. It should be understood that various aspects of the application can be implemented alone or in combination with one or more other aspects. It should also be understood that the description and specific examples herein are intended to be illustrative only and are not intended to limit the scope of the present application.

[0120] The above detailed description of several embodiments of the disclosure has been presented for the purposes of illustration and description. It is apparent to those skilled in the art that various modifications and variations can be made to the embodiments of the disclosure without departing from the spirit and scope of the disclosure. The scope of protection of the disclosure is defined by the appended claims.

Claims

1. A task planning optimization method for emergency material supply and transportation guarantee, characterized by: include: An emergency material supply model is constructed with the minimum total freight cost and the completion of each task within the task time limit as the objective function; Solve the optimal solution of the emergency material supply model and output a resource allocation plan; Build an information-task association model, associate task decision information with task stages based on historical task information, and output multi-label classification results; A multi-objective optimization model for emergency logistics tasks is constructed using the resource allocation plan as input, the multi-label classification results as decision vectors for different task stages, task cost limits and feasible paths without loops during delivery as constraints, and minimizing task risk, minimizing task time, and minimizing the number of carriers as objective functions; and The multi-objective optimization model is solved based on a multi-objective genetic algorithm, and the model output result is adjusted according to the real-time feedback of environmental parameters and task parameters to obtain the optimal delivery strategy in a real-time environment.

2. The task planning optimization method for emergency material supply and transportation guarantee according to claim 1 is characterized in that: The objective function is to minimize the total freight cost and complete each task within the task time limit. The emergency material supply model is constructed as follows: Construct the objective function of the emergency material supply model: s.t.P={p j |T ij ≤T j (j=1,2,...,n)} In the formula, Z is the total freight, m ​​is the number of material supply points, n is the number of tasks, C ij is the unit freight rate, X ij is the transportation volume from material supply point i to support object j, T j The guarantee time limit for each mission, T ij is the transportation time from material supply point i to support object j, p j is a single task, P is a set of tasks; and The constraints of the emergency material supply model are determined as follows: Where a i The amount of supplies stored at each supply point i, b j is the material demand of object j for each mission, X ij is the transportation volume from material supply point i to support object j.

3. The task planning optimization method for emergency material supply and transportation guarantee according to claim 1 is characterized in that: The step of solving the optimal solution of the emergency material supply model and outputting a resource allocation plan includes: Calculate the unit freight rate and transportation time for each transportation route, and select the transportation route with the lowest unit freight rate and the shortest transportation time as the initial configuration plan for material allocation until all materials and needs are met, completing the initial configuration plan; Performing optimal judgment on the initial configuration scheme based on the test numbers of all non-basic variables, and if all the test numbers are not less than 0, the current scheme is the optimal solution; and If there is a test number less than 0, the current solution is not the optimal solution. Then the solution is adjusted and the optimal judgment is continued until all the test numbers are greater than 0, and the optimal solution of the objective function of the emergency material supply model is obtained, and the resource allocation plan is output.

4. The task planning optimization method for emergency material supply and transportation guarantee according to claim 3 is characterized in that: If there is a test number less than 0, the current solution is not the optimal solution. Then the solution is adjusted and the optimal judgment is continued until all the test numbers are greater than 0. The optimal solution of the objective function of the emergency material supply model is obtained, which includes: Step 1: Select the variable corresponding to the minimum negative test number as the entry variable, compare the ratio of the variables in each constraint condition, and select the variable with the smallest ratio as the exit variable; Step 2: increase the value of the input variable, adjust the output variable to 0, and recalculate the test number of each variable; and Step 3: If there is still a test number less than 0, repeat steps 1 to 2 until all test numbers are greater than 0, and the optimal solution of the objective function of the emergency material supply model is obtained.

5. The task planning optimization method for emergency material supply and transportation guarantee according to claim 1 is characterized in that: The information-task association model is constructed to associate task decision information with task stages based on historical task information, and output multi-label classification results including: Constructing a dataset containing task decision information and task stages based on historical task information, and using the dataset as training data for a multi-label learning algorithm, wherein each data sample contains a feature vector of the task decision information and a label vector of the task stage; Using any one of a k-nearest neighbor-based multi-label classification algorithm, a kernel-based multi-label classification algorithm, and a deep semantic matching multi-label classification algorithm to construct an information-task association model, and training the information-task association model based on the training data to obtain a trained information-task association model; and The real-time task information features are input into the trained information-task association model, and a multi-label classification result is output.

6. The task planning optimization method for emergency material supply and transportation guarantee according to claim 1 is characterized in that: The multi-objective optimization model for emergency logistics tasks is constructed by taking the resource allocation plan as input, the multi-label classification results as decision vectors for different task stages, the task cost limit and the feasible path without loops during the delivery process as constraints, and minimizing the task risk, minimizing the task time, and minimizing the number of carriers as objective functions. The model includes: Construct the expression of the multi-objective optimization model: miny=[f1(x),f2(x),f3(x),...,f M (x)] Where x=(x1,x2...,x n )∈D is the decision vector formed by the multi-label learning classification results, D is the decision space formed by the decision vector, y=(f1(x), f2(x), ..., f M )∈Y represents the target space formed by the objective functions of the multi-objective optimization model, wherein the objective functions of the multi-objective optimization model include minimizing mission risk, minimizing mission time, and minimizing the number of carriers; and Determine the task cost limit constraints of the multi-objective optimization model: Where C all is the sum of the costs of all routes in the task network; C k is the cost of a single line; is the travel cost of the kth route, is the cost per unit time, for the time of delay; is the cost of safety factors; toll k It is the additional cost caused by unpredictable factors.

7. The task planning optimization method for emergency material supply and transportation guarantee according to claim 1 is characterized in that: Solving the multi-objective optimization model based on a multi-objective genetic algorithm and adjusting the model output results according to real-time feedback of environmental parameters and task parameters to obtain the optimal delivery strategy in a real-time environment includes: generating an initial population of a genetic algorithm based on the multi-label classification results, environmental information, and echelon position information output by the information-task association model; Calculating the fitness value of each individual in the initial population through the objective function of the multi-objective optimization model based on the Pareto optimization method; According to the fitness value of each individual, the individuals in the population are selected, crossed and mutated; Adjusting the decision variables and objective function of the multi-objective optimization model according to real-time environment parameters and task parameters; and The Pareto frontier is updated according to the adjusted decision variables and objective function, and the solution of the multi-objective optimization model is iteratively optimized until the maximum number of iterations is reached, and the optimal delivery strategy in a real-time environment is output.

8. The task planning optimization method for emergency material supply and transportation guarantee according to claim 7 is characterized in that: Each individual in the initial population includes task allocation, route selection and transportation tool selection. The environmental parameters include weather conditions, traffic conditions, resource availability, and emergencies. The task parameters include task priority and time limit.

9. A task planning optimization device for emergency material supply and transportation guarantee, characterized in that: The device comprises: at least one processor; and at least one memory storing a computer program; Wherein, when the computer program is executed by the at least one processor, the device executes the steps of the task planning optimization method for emergency material supply and transportation guarantee according to any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that: When executed by a processor, the computer program implements the steps of the task planning optimization method for emergency material supply and transportation guarantee according to any one of claims 1 to 8.

Images