An emergency supply chain network design method, system and electronic device

Abstract

The application provides an emergency supply chain network design method, system and electronic equipment, and relates to the technical field of emergency logistics and supply chain management. The method comprises the following steps: acquiring parameter data of a conventional supply point, a temporary distribution center, a disaster point and a backup supply point in a disaster area and constructing an initial emergency supply chain network; generating an original interruption scenario according to the number of conventional supply point interruptions and performing clustering reduction to obtain a representative scenario set; based on the representative conventional supply point interruption scenario set and the initial emergency supply chain network, optimizing pre-disaster resource allocation decisions and post-disaster emergency material allocation decisions under each representative conventional supply point interruption scenario to obtain an emergency supply chain network design scheme corresponding to the disaster area. The application fully excavates the supply capacity and spatial distribution information of the conventional supply point through clustering reduction, and is suitable for emergency material reserve layout and post-disaster allocation decisions under the conditions of multiple materials, multiple disaster points and multiple conventional supply point interruption scenarios.

Description

Technical Field

[0001] This application relates to the field of emergency logistics and supply chain management technology, and more specifically, to an emergency supply chain network design method, system, and electronic device. Background Technology

[0002] Emergency supply chains are a crucial material foundation for life-saving and basic livelihood security after disasters. Following major emergencies such as earthquakes, floods, typhoons, and public health emergencies, affected areas often experience a rapid surge in demand for emergency supplies such as drinking water, convenience foods, and tents. Simultaneously, disasters can damage supply points, disrupt roads, reduce warehousing capacity, or delay delivery responses, placing the emergency supply chain in a highly uncertain and time-sensitive operating environment.

[0003] Existing emergency supply chain network designs typically focus on issues such as facility location, material reserves, and distribution route planning, which can improve the efficiency of post-disaster material allocation to some extent. However, existing design methods do not fully utilize the supply capacity and spatial distribution information of supply points in disaster scenarios, resulting in poor adaptability of emergency supply chain network designs generated by existing methods to actual disaster scenarios. Summary of the Invention

[0004] To address the problem that existing emergency supply chain network design schemes have poor adaptability to real-world disaster scenarios, this application proposes an emergency supply chain network design method, system, and electronic equipment.

[0005] According to one aspect of this application, an emergency supply chain network design method is provided, comprising the following steps: acquiring parameter data of each node within a disaster area, wherein the nodes include regular supply points, temporary distribution centers, disaster-stricken areas, and backup supply points; the parameter data includes: the location of each node, the supply capacity or storage capacity of various emergency supplies, the transportation distance and transportation time between each node, and the emergency supply demand of the disaster-stricken areas; constructing an initial emergency supply chain network based on the parameter data, wherein the initial emergency supply chain network includes regular supply logistics connections from regular supply points to temporary distribution centers, distribution logistics connections from temporary distribution centers to disaster-stricken areas, and direct supplementary distribution logistics connections from backup supply points to disaster-stricken areas; generating all original regular supply point interruption scenarios based on the number of interruptions at regular supply points, and clustering and reducing all the original regular supply point interruption scenarios to obtain a representative set of regular supply point interruption scenarios; and optimizing pre-disaster resource allocation decisions and post-disaster emergency supply allocation decisions under representative regular supply point interruption scenarios based on the representative set of regular supply point interruption scenarios and the initial emergency supply chain network to obtain an emergency supply chain network design scheme corresponding to the disaster area.

[0006] Based on the above aspects, in one embodiment, based on the representative set of regular supply point interruption scenarios and the initial emergency supply chain network, the pre-disaster resource allocation decision and the post-disaster emergency material allocation decision under each representative regular supply point interruption scenario are optimized to obtain the emergency supply chain network design scheme corresponding to the disaster area. This includes: solving a bi-objective two-stage stochastic mixed integer programming model based on the representative set of regular supply point interruption scenarios and the initial emergency supply chain network to obtain the optimal solution for the decision variables; the bi-objective two-stage stochastic mixed integer programming model has a first-stage variable for characterizing the pre-disaster resource allocation decision and a second-stage variable for characterizing the pre-disaster resource allocation decision. The second-stage variables characterize emergency material allocation decisions under scenarios where representative conventional supply points are disrupted after a disaster. The bi-objective two-stage stochastic mixed integer programming model uses minimizing economic costs and minimizing comprehensive deprivation costs as dual optimization objectives. The pre-disaster resource allocation decisions include the activation decisions of conventional supply points, temporary distribution centers, and backup supply points; the establishment decisions of cooperative relationships between conventional supply points and temporary distribution centers, between temporary distribution centers and disaster-stricken areas, and between backup supply points and disaster-stricken areas; and the decisions on the safety stock allocation of various emergency materials within temporary distribution centers. Based on the optimal solutions of the decision variables, the emergency supply chain network design scheme is determined.

[0007] According to the above aspects, in one embodiment, the step of solving a bi-objective two-stage stochastic mixed integer programming model based on the representative set of regular supply point interruption scenarios and the initial emergency supply chain network to obtain the optimal solution for the decision variables includes: S1, solving the model with minimizing economic cost and minimizing comprehensive deprivation cost as single objectives respectively, obtaining two anchor solutions, and determining the value range of the comprehensive deprivation cost objective based on the two anchor solutions; S2, setting multiple ε value points within the value range, transforming the comprehensive deprivation cost objective into an ε constraint, and solving the model with minimizing economic cost as the main objective at each ε value point to obtain the corresponding candidate solution; S3, for all The candidate solutions are subjected to non-dominated screening, and dominated solutions are deleted to form an initial Pareto non-dominated solution set; S4, based on the spacing distribution of adjacent solutions in the target space in the initial Pareto non-dominated solution set, sparse target space regions are identified, and new ε-value points are inserted in the ε interval corresponding to the target space region. The corresponding ε-constrained subproblems are solved again, and the new solutions are added to the candidate solutions; S5, the candidate solutions after adding the new solutions are subjected to non-dominated screening again, and the Pareto non-dominated solution set is updated; S6, steps S4 to S5 are repeated until the preset termination condition is met, and the target Pareto non-dominated solution set is obtained as the optimal solution of the decision variable.

[0008] According to the above aspects, in one embodiment, determining the emergency supply chain network design scheme based on the optimal solution of decision variables includes: normalizing the economic cost and comprehensive deprivation cost corresponding to each non-dominated solution in the target Pareto non-dominated solution set to obtain the normalized economic cost and normalized comprehensive deprivation cost corresponding to each non-dominated solution; obtaining the decision weight combination corresponding to various decision preferences, wherein the decision preferences include the lowest economic cost preference, the partial economic cost preference, the equilibrium compromise preference, the partial deprivation cost preference, and the lowest deprivation cost preference; for each decision preference, calculating the comprehensive evaluation value of each non-dominated solution corresponding to the decision preference based on the corresponding decision weight combination, the normalized economic cost and the normalized comprehensive deprivation cost corresponding to each non-dominated solution; for each decision preference, determining the non-dominated solution with the smallest comprehensive evaluation value among the non-dominated solutions corresponding to the decision preference as the optimal solution corresponding to the decision preference; and determining the emergency supply chain network design scheme based on the optimal solutions corresponding to various decision preferences; wherein the emergency supply chain network design scheme includes one or more of the lowest economic cost solution, the partial economic cost solution, the equilibrium compromise solution, the partial deprivation cost solution, and the lowest deprivation cost solution.

[0009] According to the above aspects, in one embodiment, constructing an initial emergency supply chain network based on the parameter data includes: constructing the initial emergency supply chain network according to a combination of resilience strategies and the parameter data, wherein the combination of resilience strategies includes: the temporary distribution center establishing regular supply logistics connections with at least two regular supply points when activated; pre-setting safety stock of various emergency supplies in each of the temporary distribution centers, the configuration amount of the safety stock being determined by the safety stock configuration amount in the first stage variables; and the backup supply points establishing cooperative relationships with disaster-stricken areas before the disaster and being triggered to activate when regular supply capacity is insufficient after the disaster, directly delivering emergency supplies to the disaster-stricken areas.

[0010] According to the above aspects, in one embodiment, the representative set of regular supply point interruption scenarios includes at least one of the representative regular supply point interruption scenarios; the step of clustering and reducing all the original regular supply point interruption scenarios to obtain the representative set of regular supply point interruption scenarios includes: constructing a feature vector to characterize each original regular supply point interruption scenario; standardizing the feature vectors of each original regular supply point interruption scenario and using a clustering algorithm to divide all the original regular supply point interruption scenarios into several classes; selecting the original regular supply point interruption scenarios corresponding to the cluster centers of each class as candidate regular supply point interruption scenarios; and merging the original regular supply point interruption scenarios belonging to the same candidate regular supply point interruption scenario to obtain at least one representative regular supply point interruption scenario.

[0011] Based on the above aspects, in one embodiment, constructing a feature vector to characterize each original regular supply point interruption scenario includes: calculating a comprehensive loss score based on the supply capacity loss of various emergency supplies and preset material importance weights under each original regular supply point interruption scenario; calculating the spatial centroid coordinates and the average discrete distance of each interrupted regular supply point relative to the spatial centroid coordinates based on the geographical coordinates of all interrupted regular supply points under each original regular supply point interruption scenario; and combining the comprehensive loss score, spatial centroid coordinates, and average discrete distance to form the space-capacity joint feature vector.

[0012] Based on the above aspects, in one embodiment, the economic cost includes pre-disaster costs and post-disaster expected costs; the pre-disaster costs include fixed activation costs of regular supply points, fixed activation costs of temporary distribution centers, fixed activation costs of backup supply points, costs of establishing cooperative relationships between nodes, safety stock contract costs, safety stock procurement costs, safety stock transportation costs, and safety stock holding costs; the post-disaster expected costs include emergency material procurement costs of regular supply channels and backup supply channels under different representative regular supply point interruption scenarios, as well as transportation costs in the process of regular supply, safety stock mobilization, and backup supply; and / or, the comprehensive deprivation cost includes absolute deprivation costs of disaster-stricken points and relative deprivation costs between disaster-stricken points due to differences in protection levels; wherein, the absolute deprivation cost is calculated based on the unmet demand and maximum scarcity time of disaster-stricken points using an exponential unit deprivation cost function; the maximum scarcity time is determined based on the maximum transportation time of all feasible transportation paths in the emergency supply chain network.

[0013] According to another aspect of this application, an emergency supply chain network design system is provided, comprising: a data acquisition module for acquiring parameter data of each node within a disaster area, wherein the nodes include regular supply points, temporary distribution centers, disaster-stricken areas, and backup supply points; the parameter data includes: the location of each node, the supply capacity or storage capacity of various emergency supplies, the transportation distance and transportation time between each node, and the emergency supply demand of the disaster-stricken area; and a network construction module for constructing an initial emergency supply chain network based on the parameter data, wherein the initial emergency supply chain network includes regular supply logistics connections from regular supply points to temporary distribution centers, and connections from temporary distribution centers to disaster-stricken areas. The system includes: a distribution logistics connection at disaster sites, and a direct supplementary distribution logistics connection from backup supply points to disaster sites; a scenario generation and reduction module, used to generate all original regular supply point interruption scenarios based on the number of interruptions at regular supply points, and to cluster and reduce all the original regular supply point interruption scenarios to obtain a representative set of regular supply point interruption scenarios; and an optimization and solution module, used to optimize pre-disaster resource allocation decisions and post-disaster emergency material allocation decisions under each representative regular supply point interruption scenario based on the representative set of regular supply point interruption scenarios and the initial emergency supply chain network, to obtain an emergency supply chain network design scheme corresponding to the disaster area.

[0014] According to another aspect of this application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the above-described method when executing the computer program.

[0015] This application constructs an initial emergency supply chain network containing three types of logistics connections based on parameter data of regular supply points, temporary distribution centers, disaster-stricken points, and backup supply points within the disaster area. By generating all original regular supply point interruption scenarios and performing clustering reduction, the clustering process simultaneously merges scenarios based on the supply capacity loss information and spatial distribution information of regular supply points. The resulting representative scenario set retains both the supply capacity differences under different interruption scenarios and the spatial distribution characteristics of the interruption points. This allows the final network design scheme to adapt to changes in supply capacity and spatial distribution during actual disasters, effectively improving the adaptability of the scheme. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating an emergency supply chain network design method according to an embodiment of this application.

[0017] Figure 2 This is a schematic diagram of an emergency supply chain network structure according to an embodiment of this application.

[0018] Figure 3This is a schematic diagram of the structure of an emergency supply chain network design system according to an embodiment of this application.

[0019] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application.

[0020] Figure 5 This is a schematic diagram of a Pareto non-dominated solution set according to an embodiment of this application. Detailed Implementation

[0021] Figure 1 This is a flowchart illustrating an emergency supply chain network design method according to an embodiment of this application. Figure 1 As shown, the emergency supply chain network design method provided in this embodiment includes the following steps S101 to S104.

[0022] Step S101: Obtain parameter data for each node within the disaster area. The nodes include regular supply points, temporary distribution centers, disaster-stricken areas, and backup supply points. The parameter data includes: the location of each node, the supply capacity or storage capacity of various emergency supplies, the transportation distance and time between nodes, and the emergency supply demand at the disaster-stricken areas.

[0023] In one possible implementation, node location data is sourced from an electronic map platform and uniformly converted into latitude and longitude coordinates; subsequently, the transportation distance and time between each node are calculated based on these coordinates. Emergency supplies include, but are not limited to, drinking water, convenience food, and tents.

[0024] Among them, regular supply points are used to provide emergency supplies through regular channels to temporary distribution centers before and after a disaster; temporary distribution centers are used to receive supplies delivered by regular supply points, set up safety stock, and deliver emergency supplies to disaster-stricken areas after a disaster; disaster-stricken areas are used to characterize areas where there is a demand for supplies after a disaster occurs; and backup supply points are used to bypass temporary distribution centers and directly replenish emergency supplies to disaster-stricken areas when regular supply capacity is insufficient.

[0025] Step S102: Construct an initial emergency supply chain network based on parameter data. The initial emergency supply chain network includes regular supply logistics connections from regular supply points to temporary distribution centers, distribution logistics connections from temporary distribution centers to disaster-stricken areas, and direct supplementary distribution logistics connections from backup supply points to disaster-stricken areas. Figure 2 An example of the initial emergency supply chain network of this application is shown.

[0026] Step S103: Based on the number of interruptions at regular supply points, generate all original regular supply point interruption scenarios, and cluster and reduce all original regular supply point interruption scenarios to obtain a representative set of regular supply point interruption scenarios.

[0027] In one possible implementation, the interruption scenarios are divided into different interruption orders based on the number of regular supply point interruptions (e.g., no regular supply point interruption, one regular supply point interruption, two regular supply point interruptions, and three regular supply point interruptions). All original regular supply point interruption scenarios for each interruption order are generated through enumeration. The occurrence probability corresponding to each interruption order is given by historical statistical data, risk assessment results, or decision-makers. Subsequently, a clustering algorithm is used to cluster and reduce all original regular supply point interruption scenarios, resulting in a representative set of regular supply point interruption scenarios and their occurrence probabilities.

[0028] Step S104: Based on the set of representative regular supply point interruption scenarios and the initial emergency supply chain network, optimize the pre-disaster resource allocation decision and the emergency material allocation decision under the representative regular supply point interruption scenarios after the disaster, and obtain the emergency supply chain network design scheme corresponding to the disaster area.

[0029] In one possible implementation, the optimization can be achieved in the following way: First, an optimization model is constructed based on a set of representative scenarios of disruption at regular supply points and an initial emergency supply chain network. The decision variables in this optimization model include pre-disaster resource allocation decision variables and post-disaster emergency material allocation decision variables under various scenarios.

[0030] Secondly, the optimization model is solved to obtain the optimal solution for the decision variables.

[0031] Finally, based on the optimal solutions of the decision variables, the emergency supply chain network design scheme is determined. The values ​​of the pre-disaster resource allocation decision variables in the optimal solutions correspond to the pre-disaster network design scheme, including which regular supply points, temporary distribution centers, and backup supply points are activated, which cooperative relationships are established, and how much safety stock is allocated to each temporary distribution center; the values ​​of the post-disaster emergency material allocation decision variables correspond to the material allocation scheme under the scenario of interruption of each representative regular supply point.

[0032] According to this embodiment, an initial emergency supply chain network containing three types of logistics connections is constructed based on the parameter data of each node in the disaster area. By generating all original regular supply point interruption scenarios and performing clustering reduction, the system systematically covers all situations under various regular supply point interruption quantities and combinations, enabling the representative scenario set to reflect the actual possible combinations of regular supply point interruptions and their characteristics, effectively improving the adaptability of the final generated network design scheme in actual disaster scenarios. By optimizing pre-disaster resource allocation and post-disaster material allocation, pre-disaster decisions can be optimized based on various possible post-disaster interruption scenarios, and the final output network design scheme can better adapt to actual disaster scenarios.

[0033] According to one embodiment of this application, based on a representative set of regular supply point disruption scenarios and an initial emergency supply chain network, the pre-disaster resource allocation decision and the post-disaster emergency material allocation decision under various representative regular supply point disruption scenarios are optimized, including the following steps S201 to S202: Step S201: Based on a set of representative routine supply point disruption scenarios and an initial emergency supply chain network, solve a bi-objective two-stage stochastic mixed-integer programming model. The bi-objective two-stage stochastic mixed-integer programming model has first-stage variables characterizing pre-disaster resource allocation decisions and second-stage variables characterizing emergency material allocation decisions under post-disaster disruption scenarios at representative routine supply points. The bi-objective two-stage stochastic mixed-integer programming model has two optimization objectives: minimizing economic cost and minimizing overall deprivation cost.

[0034] Among them, pre-disaster resource allocation decisions include decisions on activating regular supply points, temporary distribution centers, and backup supply points; decisions on establishing cooperative relationships between regular supply points and temporary distribution centers, between temporary distribution centers and disaster-stricken areas, and between backup supply points and disaster-stricken areas; and decisions on the safety stock allocation of various emergency supplies within temporary distribution centers.

[0035] The model is solved using a multi-objective optimization method to obtain the optimal solution for the decision variables.

[0036] Step S202: Determine the emergency supply chain network design scheme based on the optimal solution of the decision variables.

[0037] Based on the optimal solutions of the decision variables, an emergency supply chain network design scheme is determined. For example, the values ​​of the first-stage variables in the optimal solutions of the decision variables correspond to the pre-disaster network design scheme, and the values ​​of the second-stage variables correspond to the material allocation scheme under the scenario of interruption of each representative conventional supply point.

[0038] In this embodiment, the bi-objective two-stage stochastic mixed integer programming model used in this application is described. The sets involved in this model include: the set of regular supply points I, the set of temporary distribution centers R, the set of disaster-stricken points P, the set of backup supply points L, the set of emergency material types K, and the set of representative interruption scenarios S.

[0039] I. Economic Cost Objective Function: Economic costs include pre-disaster costs and expected post-disaster costs.

[0040] According to one embodiment of this application, pre-disaster costs are the sum of various costs incurred in a one-time manner during the pre-disaster phase, including: (i) fixed activation costs of regular supply points, i.e., the fixed construction or operating costs required to determine which regular supply points to activate before a disaster. As the core node of the regular supply system, the activation decision of regular supply points directly determines the foundation of post-disaster material supply capacity; (ii) fixed activation costs of temporary distribution centers, i.e., the fixed construction or operating costs required to determine which temporary distribution centers to activate before a disaster. As a key node for material transfer and storage, the location and activation decisions of temporary distribution centers directly affect the efficiency and coverage of post-disaster material distribution; (iii) fixed activation costs of backup supply points, i.e., the fixed cooperation or operating costs required to determine which backup supply points to activate before a disaster. As an emergency redundancy resource, the activation decision of backup supply points determines the availability of post-disaster external compensation; (iv) costs of establishing cooperative relationships between nodes, i.e., the costs required to establish logistics cooperation relationships between nodes before a disaster. These include: the cost of establishing cooperative relationships between regular supply points and temporary distribution centers, the cost of establishing cooperative relationships between temporary distribution centers and disaster-stricken areas, and the cost of establishing cooperative relationships between backup supply points and disaster-stricken areas. The above cooperative relationships constitute a complete emergency supplies distribution network; (v) Safety stock contract costs, which are the fixed costs required to sign safety stock contracts with regular supply points before a disaster. After signing the contract, regular supply points can distribute safety stock to temporary distribution centers before a disaster; (vi) Safety stock procurement costs, which are the procurement costs of emergency supplies purchased and stored in temporary distribution centers before a disaster. This cost depends on the pre-disaster purchase unit price and the safety stock allocation quantity; (vii) Safety stock transportation costs and holding costs. Safety stock transportation costs refer to the transportation costs of transporting emergency supplies from regular supply points to temporary distribution centers. This cost depends on the transportation distance, the unit distance transportation cost of the supplies, and the safety stock allocation quantity. Safety stock holding costs refer to the warehousing management costs incurred during the storage of supplies in temporary distribution centers.

[0041] According to one embodiment of this application, the expected post-disaster cost is the expected value of the allocation cost under the scenario of disruption of representative conventional supply points after a disaster, including: (i) the emergency material procurement cost of conventional supply channels, i.e., the cost of procuring emergency materials through conventional supply points after a disaster. This cost depends on the unit price of conventional procurement and the volume of conventional supply distribution after a disaster; (ii) transportation costs, including three parts: transportation costs of conventional supply channels, transportation costs of safety stock mobilization and distribution, and transportation costs of backup supply channels. The transportation cost of conventional supply channels is the total transportation cost of materials from conventional supply points to temporary distribution centers and then to disaster-stricken areas. The transportation cost of safety stock mobilization and distribution is the transportation cost of materials from temporary distribution centers to disaster-stricken areas. The transportation cost of backup supply channels is the transportation cost of materials directly from backup supply points to disaster-stricken areas; (iii) the emergency material procurement cost of backup supply channels, i.e., the cost of procuring emergency materials through backup supply points after a disaster. This cost only occurs when conventional supply capacity is insufficient. The post-disaster allocation cost under each scenario is multiplied by the probability of occurrence of that scenario. The summation afterwards yields the expected post-disaster cost.

[0042] The sum of pre-disaster costs and expected post-disaster costs constitutes the total economic cost. This application, through comprehensive quantification of economic costs, enables the model to accurately balance the pre-disaster construction investment and post-disaster operating costs during the optimization process, avoiding insufficient post-disaster support capacity due to excessive pursuit of low pre-disaster costs, or reduced economic efficiency due to over-allocation of resources.

[0043] For example, pre-disaster costs The calculation formula is:

[0044] Where i is the index of regular supply points, r is the index of temporary distribution centers, l is the index of backup supply points, p is the index of disaster-stricken points, and k is the index of emergency supplies types. To enable the fixed operating costs of regular supply point i, The enable variable for regular supply point i (1 for enabling regular supply point i, 0 otherwise); To cover the fixed operating costs of activating the temporary distribution center r, The variable for enabling temporary distribution center r (r is 1 if enabling temporary distribution center, otherwise it is 0); The fixed operating costs of activating the backup supply point l The enable variable for the backup supply point l (1 for enable backup supply point l, 0 otherwise). The cost of establishing a partnership between a regular supply point i and a temporary distribution center r. The variable represents the cooperative relationship between regular supply point i and temporary distribution center r (1 for establishing a cooperative relationship, 0 otherwise). The cost of establishing a partnership between the temporary distribution center r and the disaster-stricken site p. The variable represents the cooperative relationship between the temporary distribution center r and the disaster-stricken point p (1 if a cooperative relationship is established, 0 otherwise). The cost of establishing a cooperative relationship between backup supply point l and disaster-stricken point p. The variable represents the cooperative relationship between the backup supply point l and the disaster-stricken point p (1 if a cooperative relationship is established, 0 otherwise). The fixed costs of signing a safety stock contract for regular supply point i, This is the contract signing variable for regular supply point i (1 for a signed contract, 0 otherwise). Let k be the unit price of the supplies purchased before the disaster. Let k be the unit distance transportation cost of material. The distance from the regular supply point i to the temporary distribution center r. Let K be the unit inventory holding cost of material k in the temporary distribution center. This indicates the safety stock of supplies k that were previously stored at regular supply point i and transferred to temporary distribution center r before the disaster.

[0045] For example, the deployment cost in post-disaster scenario s The calculation formula is:

[0046] in, The unit price of supplies k procured from regular supply points after the disaster. The distance from the temporary distribution center r to the disaster-stricken point p. is a non-negative real variable, representing the quantity of supplies k (excluding safety stock) delivered from regular supply point i to disaster point p via temporary distribution center r under scenario s. Let be a non-negative real variable, representing the quantity of supplies k that the temporary distribution center r uses to deliver to the disaster-stricken point p under scenario s. Let k be the unit distance transportation cost of material. The distance from the temporary distribution center r to the disaster-stricken point p. The unit price of supplies k procured from backup supply points after the disaster. The distance from backup supply point l to disaster-stricken point p. Let be a non-negative real variable, representing the quantity of supplies k delivered from backup supply point l to disaster-stricken point p under scenario s.

[0047] For example, the formula for calculating the economic cost target is:

[0048] in, Let be the probability of scenario s occurring.

[0049] II. Overall Exemption Cost Objective Function: In one embodiment of this application, the comprehensive deprivation cost includes the absolute deprivation cost of the disaster-affected site and the relative deprivation cost between disaster-affected sites due to differences in protection levels.

[0050] Absolute deprivation cost is used to quantify the direct welfare loss suffered by disaster-stricken groups due to the lack of emergency supplies. Its core idea is that the longer the shortage of supplies lasts, the greater the suffering and loss for the affected groups, and this loss increases non-linearly and rapidly over time. To characterize the welfare loss suffered by disaster-stricken groups due to delayed arrival of supplies and unmet needs, this application adopts an exponential unit deprivation cost function:

[0051] in, This represents the unit deprivation cost of resource k during a scarcity period of t. is the proportional parameter of the deprivation cost function for material k (controlling the initial level of deprivation cost). Let be the growth rate parameter of the deprivation cost function for resource k (controlling the rate at which deprivation cost increases over time), where t represents the duration for which the demand for resource k at the disaster-stricken point is not met, i.e., the scarcity time, and e is the base of the natural logarithm. This function has the following characteristics: when the scarcity time t is small, the unit deprivation cost increases relatively slowly; as the scarcity time t increases, the rate of increase in unit deprivation cost accelerates, reflecting the objective law that "the longer the wait, the greater the increase in suffering for every additional minute."

[0052] For example, the formula for calculating the absolute cost of depriving resource k from disaster point p in scenario s is as follows:

[0053] in, This represents the quantity (excluding safety stock) of supplies k delivered from regular supply point i to disaster-stricken point p via temporary distribution center r under scenario s. The transportation time from the regular supply point i to the disaster-stricken point p via the temporary distribution center r. In scenario s, the temporary distribution center r uses safety stock to deliver the quantity of supplies k to the disaster-stricken point p. The transportation time from the temporary distribution center r to the disaster-stricken point p. Let k be the quantity of supplies k delivered from backup supply point l to disaster-stricken point p under scenario s. The transportation time from backup supply point l to disaster-stricken point p. Let T be the unmet demand of disaster point p for material k under scenario s, and T be the maximum scarcity time, which is the maximum transportation time of all feasible transportation routes in the emergency supply chain network.

[0054] This calculation method links absolute deprivation cost to two factors: first, the scale of unmet demand (the larger the scale, the higher the deprivation cost); and second, the duration of scarcity, T (the larger T is, the higher the unit deprivation cost). This allows for a simultaneous characterization of both the "quantity" and "time" dimensions of resource scarcity.

[0055] Relative deprivation cost is used to quantify the psychological and social welfare losses caused by differences in the level of material support between different disaster-stricken areas. Its core idea is that even if two disaster-stricken areas receive a certain amount of supplies, if the level of support received by one disaster-stricken area is significantly lower than that of the other disaster-stricken areas, the disaster-stricken population in that disaster-stricken area will experience additional welfare losses due to "relative deprivation".

[0056] For example, in scenario s, the relative deprivation costs of resources k for disaster-stricken points p and j. The calculation formula is:

[0057] in, Let $\frac{p}{s}$ be the absolute cost of depriving disaster-stricken point $p of resources $k$ under scenario $s$. Let be the absolute deprivation cost of supplies k by disaster point j under scenario s. The absolute value of the difference between the two absolute deprivation costs reflects the difference in welfare loss caused by the difference in protection level between the two disaster points. When the difference in absolute deprivation costs between the two disaster points is greater, the disaster group with the lower protection level will feel a stronger sense of relative deprivation.

[0058] For example, the formula for calculating the overall deprivation cost target is as follows:

[0059] in, The probability of the scenario occurring. This represents the relative deprivation cost coefficient, where d represents the number of pairs of disaster-affected points. Let be the difference in relative deprivation of resources k between disaster-affected point p and disaster-affected point j in scenario s. The objective function applies probability weights to the sum of absolute deprivation costs and relative deprivation costs in each scenario, where the relative deprivation cost component is normalized (divided by d) to eliminate the influence of the number of disaster-affected points on the magnitude of the relative deprivation cost.

[0060] According to this embodiment, by introducing absolute deprivation cost, the model can consider the cumulative effect of material scarcity over time in the optimization process, guiding the system to prioritize the timeliness of material delivery; by introducing relative deprivation cost, the model can take into account the fairness of material allocation among disaster-stricken areas in the optimization process, avoiding the imbalance where some disaster-stricken areas have sufficient materials while others suffer from severe shortages; the comprehensive deprivation cost target formed by the combination of the two makes the optimization results more in line with the humanitarian requirements of disaster relief.

[0061] In some embodiments, the bi-objective two-stage stochastic mixed integer programming model further includes at least one of the following constraints.

[0062] 1) Coverage constraints of disaster-affected areas:

[0063] in, Let r be the variable representing the cooperative relationship between the temporary distribution center r and the disaster-stricken point p. This constraint ensures that each disaster-stricken point is served by at least one temporary distribution center.

[0064] 2) Logical constraints on the cooperative relationship between regular supply points and temporary distribution centers:

[0065]

[0066] in, For enabling variables of regular supply point i, The enabling variable for temporary distribution center r. This variable represents the cooperative relationship between the two entities. This constraint ensures that a cooperative relationship can only be established if both regular supply points and temporary distribution centers are in operation.

[0067] 3) Logical constraints on the cooperative relationship between temporary distribution centers and disaster-stricken areas:

[0068] in, The enabling variable for temporary distribution center r. Let r be the variable representing the cooperative relationship between the temporary distribution center r and the disaster-stricken point p. This constraint ensures that only activated temporary distribution centers can provide services to the disaster-stricken point.

[0069] 4) Temporary distribution center capacity constraints:

[0070]

[0071] in, This refers to the safety stock of supplies k stored at regular supply points i and temporary distribution centers r before a disaster. The maximum capacity of temporary distribution center r for material k. The enabling variable for temporary distribution center r. Let K be the quantity of supplies (excluding safety stock) delivered from regular supply point i to disaster-stricken point p via temporary distribution center r under scenario s. This constraint limits the pre-disaster inventory configuration and post-disaster turnover to the capacity of the temporary distribution center.

[0072] 5) Safety stock redeployment constraints:

[0073]

[0074] in, This represents the quantity of supplies k that the temporary distribution center r uses to deliver to disaster-stricken point p from its safety stock in scenario s. This indicates the safety stock of supplies k that were previously stored at the regular supply point i and transferred to the temporary distribution center r before the disaster. Let K be the total demand of emergency supplies k for disaster-stricken point p under scenario s. Let r be the variable representing the cooperative relationship between the temporary distribution center r and the disaster-stricken point p. This constraint ensures that the amount of safety stock mobilized after a disaster does not exceed the amount of pre-disaster stock.

[0075] 6) Constraints on available supply capacity at regular supply points:

[0076]

[0077]

[0078] in, The quantity of supplies k (excluding safety stock) delivered from regular supply point i to disaster-stricken point p via temporary distribution center r under scenario s. Let K be the total demand of emergency supplies k for disaster-stricken point p under scenario s. Let i be the variable representing the cooperative relationship between the regular supply point i and the temporary distribution center r. Let r be the variable representing the cooperative relationship between the temporary distribution center r and the disaster-stricken point p. This constraint characterizes the available supply capacity of regular supply points under different disruption scenarios.

[0079] 7) Constraints on backup supply capacity:

[0080]

[0081] in, Let k be the quantity of supplies k delivered from backup supply point l to disaster-stricken point p under scenario s. The maximum supply capacity of backup supply point l to material k. The activation variable for backup supply point l. Let k be the quantity of supplies k delivered from backup supply point l to disaster-stricken point p under scenario s. Let K be the total demand of emergency supplies k for disaster-stricken point p under scenario s. This is a variable representing the cooperative relationship between the backup supply point l and the disaster-stricken point p. This constraint limits the maximum supply from the backup supply point under contractual and capacity conditions.

[0082] 8) Constraints on the Availability of Demand:

[0083] in, Let k be the amount of supplies k actually received by disaster-stricken point p in scenario s. The quantity of supplies k (excluding safety stock) delivered from regular supply point i to disaster-stricken point p via temporary distribution center r under scenario s. This represents the quantity of supplies k that the temporary distribution center r uses to deliver to disaster-stricken point p from its safety stock in scenario s. Let k be the quantity of supplies k delivered from backup supply point l to disaster-stricken point p under scenario s. This constraint describes the quantity of supplies the disaster-stricken point obtains from three channels: regular supply, safety stock, and backup supply.

[0084] 9) Requirements constraints not met:

[0085]

[0086] in, Let K be the unmet demand for supplies k at disaster-stricken point p in scenario s. Let K be the total demand of emergency supplies k for disaster-stricken point p under scenario s. Let k be the quantity of supplies k actually received by disaster-stricken point p under scenario s. This constraint calculates the unmet demand of each disaster-stricken point under different scenarios.

[0087] 10) Minimum demand satisfaction rate constraint:

[0088] in, The probability of the scenario occurring. Let k be the amount of supplies k actually received by disaster-stricken point p in scenario s. Let K be the total demand of emergency supplies k for disaster-stricken point p under scenario s. This is the minimum demand fulfillment rate threshold for each type of emergency supplies. This constraint ensures that the weighted demand fulfillment rate for each type of emergency supplies is not lower than the preset threshold.

[0089] 11) Variable value constraints:

[0090]

[0091]

[0092] This constraint limits the range of values ​​for all decision variables: inventory, distribution volume, and shortage are all physical quantities and can only take the values ​​0 or positive integers; costs and losses cannot have negative values, and the minimum deprivation loss is 0; there are only two states: whether the facility is activated or whether the cooperative relationship is established.

[0093] According to this embodiment, by constructing a two-stage stochastic mixed integer programming model with economic cost and comprehensive deprivation cost as dual objectives, the pre-disaster network design and post-disaster scenario response are considered in a unified manner, overcoming the shortcomings of traditional single-objective optimization that makes it difficult to balance cost and level of protection. By separating the variables in the first stage and the variables in the second stage, the pre-disaster resource allocation decisions are optimized based on the expected performance under various possible post-disaster interruption scenarios, realizing the integrated optimization of pre-disaster proactive defense and post-disaster rapid response. By introducing the comprehensive deprivation cost objective, the cumulative effect of material scarcity over time and the effect of unfair distribution among disaster-stricken points are simultaneously characterized, and the optimization results are more in line with the humanitarian requirements of disaster relief.

[0094] According to an embodiment of this application, the process of solving a bi-objective two-stage stochastic mixed integer programming model using the adaptive ε-constraint method to obtain the optimal solution for the decision variables includes the following steps S301 to S306: Step S301: Solve the model with the single objectives of minimizing economic cost and minimizing comprehensive deprivation cost respectively to obtain two anchor solutions, and determine the range of values ​​for the comprehensive deprivation cost objective based on the two anchor solutions.

[0095] For example, first solve the model with the single objective of minimizing economic cost (i.e., ignore the overall deprivation cost objective and only optimize the economic cost), and record the overall deprivation cost objective value corresponding to the solution. Then, solve the model with the single objective of minimizing the overall deprivation cost (i.e., ignoring the economic cost objective and only optimizing the overall deprivation cost), and record the overall deprivation cost target value corresponding to the solution. Therefore, the range of values ​​for the comprehensive deprivation cost target is determined to be [ , ].

[0096] The two anchor point solutions mentioned above represent the extreme cases when the two objectives are optimal. In practice, and This constitutes the theoretical lower and upper bounds of the comprehensive deprivation cost target.

[0097] Step S302: Set multiple values ​​within the range of values. The value point transforms the overall deprivation cost target into... Constraints, and in each The solution is obtained by minimizing economic cost at the given value point, thus yielding the corresponding candidate solutions.

[0098] For example, within the range of values ​​[ , M initial values ​​are evenly distributed within the [inner area]. The sampling points are M, which is the preset number of initial sampling points, usually set according to actual computing resources and accuracy requirements.

[0099] Transform the overall deprivation cost target into a constraint. That is, when the total cost of deprivation does not exceed Under the premise of minimizing economic cost, the solution is obtained with the sole objective of minimizing economic cost. Solve the corresponding subproblems at each value point to obtain M feasible solutions as candidate solutions.

[0100] Step S303: Perform non-dominated screening on all candidate solutions, delete dominated solutions, and form the initial Pareto non-dominated solution set.

[0101] Here, a "dominated solution" refers to a candidate solution that is not superior to the other solution in either of the two objective dimensions. For example, for any two candidate solutions A and B, if the economic cost of A is not higher than the economic cost of B, and the comprehensive deprivation cost of A is not higher than the comprehensive deprivation cost of B, and at least one objective is strictly superior to B (i.e., lower economic cost or lower comprehensive deprivation cost), then A is said to dominate B, B is a dominated solution, and should be deleted from the solution set.

[0102] After the non-dominated screening, the remaining solutions are all Pareto non-dominated solutions, that is, there is no other solution that is better than it in both objective dimensions. These solutions constitute the initial Pareto non-dominated solution set.

[0103] Step S304: Based on the spacing distribution of adjacent solutions in the target space in the initial Pareto non-dominated solution set, identify sparsely distributed target space regions, and in the corresponding target space regions... Inserting new interval Find the value point and solve the corresponding solution again. Constrain the subproblem and add the new solution to the candidate solution.

[0104] For example, the initial Pareto non-dominated solution set is sorted in ascending order of the comprehensive stripping cost objective value, and the distance between adjacent solutions along the comprehensive stripping cost dimension is calculated. When the distance between any two adjacent solutions is greater than a preset interval threshold, it is determined that the solution set distribution in that region is relatively sparse, indicating that the current... The sampling points are not dense enough in this region, which may have missed important Pareto front information, thus triggering... Interval encryption mechanism.

[0105] Under the condition of meeting encryption Within the interval (i.e., two adjacent intervals) (between values), for example, selecting one of the points as the new... Find the value point and solve the corresponding solution again. Constrain the subproblem and add the new solution to the candidate solution.

[0106] Step S305: Perform non-dominated screening again on the candidate solutions after adding the new solutions, and update the Pareto non-dominated solution set. That is, after adding the new solutions generated in step S304 to the original candidate solution set, the non-dominated screening operation is performed again, the newly added dominated solutions are deleted, and the updated Pareto non-dominated solution set is formed.

[0107] Step S306: Repeat steps S304 to S305 until the preset termination condition is met, and obtain the target Pareto non-dominated solution set as the optimal solution for the decision variable.

[0108] In one possible implementation, the preset termination condition includes at least one of the following: Condition 1: There is no encryption condition that is satisfied. Interval. That is, in the current Pareto non-dominated solution set, the distance between all adjacent solutions in the dimension of comprehensive deprivation cost is no greater than a preset threshold, indicating that the solution set distribution is sufficiently uniform.

[0109] Condition 2: Reaching the preset maximum number of encryption rounds. To avoid excessive consumption of computing resources, the preset maximum number of encryption rounds serves as the upper limit for the termination condition.

[0110] Condition 3: The preset maximum number of solutions has been reached. That is, the total number of current candidate solutions has reached the preset maximum limit.

[0111] The iteration terminates when any of the above conditions are met, and the current Pareto non-dominated solution set is output as the final result.

[0112] According to this embodiment, the range of values ​​for the comprehensive deprivation cost objective is determined by solving the single-objective model separately, ensuring complete coverage of the solution space and paving the way for subsequent... The parameter settings provide accurate theoretical boundaries; by combining initial uniform sampling with adaptive encryption, while maintaining the overall uniformity of the solution set distribution, the encryption is focused on sparse regions, making the Pareto front distribution in the target space more uniform and complete; the encryption process is carried out on sparse regions, avoiding repeated calculations in existing dense regions, and effectively controlling the solution scale while ensuring the quality of the solution set; finally, a Pareto non-dominated solution set with better coverage and more uniform distribution is obtained, providing decision-makers with a richer selection space.

[0113] According to one embodiment of this application, the process of determining an emergency supply chain network design scheme based on the optimal solution of decision variables includes the following steps S401 to S405: Step S401: Normalize the economic cost and comprehensive expropriation cost corresponding to each non-dominated solution in the target Pareto non-dominated solution set to obtain the normalized economic cost corresponding to each non-dominated solution. and normalized comprehensive deprivation cost Normalization maps the values ​​of both targets to the [0,1] interval, eliminating the dimensional differences.

[0114] Step S402: Obtain the decision weight combinations corresponding to various decision preferences. Decision preferences include the lowest economic cost preference, the preference for partial economic cost, the equilibrium compromise preference, the preference for partial deprivation cost, and the preference for lowest deprivation cost.

[0115] In one possible implementation, it is assumed that importance preferences for economic costs and comprehensive deprivation costs are assigned weights respectively. and The decision weight combination corresponding to the lowest economic cost preference is ( , = (1,0), the weight combination corresponding to the preference for economic cost is (0.75,0.25), the weight combination corresponding to the preference for equilibrium compromise is (0.5,0.5), the weight combination corresponding to the preference for deprivation cost is (0.25,0.75), and the weight combination corresponding to the preference for minimum deprivation cost is (0,1).

[0116] Step S403: For each decision preference, based on the corresponding decision weight combination, the normalized economic cost and the normalized comprehensive deprivation cost corresponding to each non-dominated solution, calculate the comprehensive evaluation value of each non-dominated solution corresponding to the decision preference.

[0117] The comprehensive evaluation value is calculated as follows:

[0118] in, .

[0119] Step S404: For each decision preference, determine the non-dominated solution with the smallest comprehensive evaluation value among all non-dominated solutions corresponding to the decision preference as the optimal solution corresponding to the decision preference.

[0120] For example, for a given combination of decision weights ( , The optimal solution n* satisfies: n*=arg F(n) in, The target Pareto non-dominated solution set.

[0121] Step S405: Based on the optimal solutions corresponding to various decision preferences, determine the emergency supply chain network design scheme. The emergency supply chain network design scheme includes one or more of the following: minimum economic cost solution, partial economic cost solution, equilibrium compromise solution, partial stripping cost solution, and minimum stripping cost solution.

[0122] In one possible implementation, the final emergency supply chain network design includes, but is not limited to, the following: a set of activated regular supply points, a set of activated temporary distribution centers, and a set of activated backup supply points; the cooperative relationship between regular supply points and temporary distribution centers, the service relationship between temporary distribution centers and disaster-stricken areas, and the cooperative relationship between backup supply points and disaster-stricken areas; the safety stock configuration of various emergency supplies in each temporary distribution center; the material distribution volume of regular supply channels, the safety stock mobilization and distribution volume, and the direct distribution volume of backup supply points under different representative scenarios of regular supply point interruption; and the evaluation results of economic costs, absolute deprivation costs, relative deprivation costs, and comprehensive deprivation costs.

[0123] According to this embodiment, by normalizing the Pareto nondominated solution, the dimensional differences between different objectives are eliminated, making the comprehensive evaluation more objective. By setting multiple decision preferences and their corresponding weight combinations, various solutions such as minimum economic cost, equilibrium compromise, and minimum deprivation cost can be output to meet the emergency support needs under different disaster scenarios and different decision preferences.

[0124] According to one embodiment of this application, constructing an initial emergency supply chain network based on parameter data includes: constructing an initial emergency supply chain network according to a combination of resilience strategies and parameter data. The combination of resilience strategies includes a multi-source procurement strategy, a safety stock strategy, and a backup supply strategy.

[0125] 1) Multi-source procurement strategy: When a temporary distribution center is put into operation, it shall establish regular supply logistics connections with at least two regular supply points.

[0126] A multi-source procurement strategy requires each activated temporary distribution center to establish supply partnerships with at least two regular supply points. This strategy prevents temporary distribution centers from becoming overly reliant on a single regular supply point. If one regular supply point experiences a disruption, the temporary distribution center can still obtain supplies from other regular supply points, thus ensuring the redundancy of supply sources.

[0127] In one possible implementation, the multi-source procurement strategy is achieved through the following multi-source procurement constraints:

[0128] in, Let be the variable representing the cooperative relationship between regular supply point i and temporary distribution center r. Let i be the activation variable for regular supply point i. This constraint means that each activated temporary distribution center must establish a partnership with at least two regular supply points.

[0129] 2) Safety stock strategy: Pre-position safety stock of various emergency supplies in each temporary distribution center. The amount of safety stock is represented by the safety stock allocation decision in the first stage variable.

[0130] The safety stock strategy involves pre-positioning some emergency supplies in temporary distribution centers before a disaster. If regular supply points are disrupted or the regular supply routes are slow to respond after a disaster, the temporary distribution centers can directly utilize the safety stock to deliver supplies to the affected areas, thereby shortening response time and reducing the deprivation losses suffered by affected groups due to waiting for supplies.

[0131] In one possible implementation, the safety stock strategy is achieved through the following constraints: Safety stock configuration constraints:

[0132]

[0133] in, This refers to the safety stock of supplies k stored at regular supply points i and temporary distribution centers r before a disaster. Let i be the maximum supply capacity of regular supply point i for material k. This represents the upper limit of the safety stock ratio. Let be the contract signing variable for regular supply point i. This constraint ensures that safety stock allocation can only occur between regular supply points with established partnerships and temporary distribution centers, and only regular supply points that have signed safety stock contracts can allocate safety stock to temporary distribution centers.

[0134] 3) Backup supply strategy: Backup supply points establish cooperative relationships with disaster-stricken areas before the disaster and are activated when regular supply capacity is insufficient after the disaster to directly deliver emergency supplies to disaster-stricken areas.

[0135] The backup supply strategy involves establishing partnerships with backup supply points before a disaster. When the supply capacity of regular supply points is insufficient or interrupted, resulting in a lack of supplies to meet demand, backup supply points can bypass temporary distribution centers and deliver supplies directly to the disaster-stricken areas, thus achieving external compensation.

[0136] In one possible implementation, the backup supply strategy is achieved through the following constraints: Logical constraints on the cooperative relationship between backup supply points and disaster-stricken areas:

[0137] in, Let l be the variable representing the cooperative relationship between the backup supply point and the disaster-stricken point p. This is the activation variable for standby supply point l. This constraint ensures that standby supply relationships can only be established by an activated standby supply point.

[0138] According to this embodiment, by simultaneously embedding three types of resilience strategies—multi-source procurement, safety stock, and backup supply—into the emergency supply chain network design, the system resilience under the condition of disruption of regular supply points can be improved from three aspects: supply source redundancy, pre-positioned reserves, and external compensation. This provides an initial emergency supply chain network with a complete structure and sufficient resilience for subsequent optimization.

[0139] According to one embodiment of this application, the representative set of regular supply point disruption scenarios includes at least one representative regular supply point disruption scenario. The process of clustering and reducing the original regular supply point disruption scenarios includes the following steps S601 to S604.

[0140] Step S601: For each original regular supply point interruption scenario, construct a feature vector to characterize that original regular supply point interruption scenario. In one possible implementation, the feature vector includes supply capacity loss features and spatial distribution features.

[0141] Step S602: Standardize the feature vectors of each original regular supply point interruption scenario, and use a clustering algorithm to divide all original regular supply point interruption scenarios into several classes.

[0142] In one possible implementation, the k-medoids clustering algorithm is used to cluster all original interruption scenes. The number of clusters is preset as a scene reduction parameter to control the balance between the number of representative scenes and the scale of subsequent model solutions. The k-medoids algorithm uses real samples from the original scenes as cluster centers, and the reduced representative scenes still come from the original scene set and have clear physical meaning.

[0143] Step S603: Select the original regular supply point interruption scenarios corresponding to the cluster centers of each type as candidate regular supply point interruption scenarios.

[0144] Since the k-medoids clustering algorithm uses real sample points as cluster centers, the candidate scenarios still come from the original scenario set. Each candidate scenario corresponds to a set of real interrupted regular supply points and their spatial distribution and supply capacity loss characteristics.

[0145] Step S604: Merge the original regular supply point interruption scenarios belonging to the same candidate regular supply point interruption scenario to obtain at least one representative regular supply point interruption scenario.

[0146] After the above clustering reduction, the original large number of interruption scenarios were grouped into a small number of representative scenarios, and each representative scenario represents a class of original scenarios with similar interruption characteristics.

[0147] According to this embodiment, by using a clustering reduction method based on scene feature vectors, a large number of original regular supply point interruption scenarios are merged into a small number of representative scenarios, which effectively reduces the solution scale of the subsequent stochastic programming model. Since the k-medoids algorithm is used with real samples as cluster centers, the reduced representative scenarios still come from the original scene set and have clear physical meaning and interpretability.

[0148] According to one embodiment of this application, the process of constructing a feature vector for each original regular supply point interruption scenario includes the following steps S701 to S703: Step S701: Calculate the comprehensive loss score based on the loss of supply capacity of various emergency supplies under the interruption scenario of each original regular supply point and the preset weight of material importance.

[0149] The comprehensive loss score is calculated as follows:

[0150] in, Let be the overall loss score for scene s. Let k be the loss amount of the k-th type of material in scenario s. This represents the total reserve of such materials. K represents the weight of the importance of the materials, and K is the set of emergency materials.

[0151] The overall loss score reflects the degree of damage to the overall supply capacity under this disruption scenario. The higher the overall loss score, the more severe the supply capacity loss caused by this scenario.

[0152] Step S702: Calculate the spatial centroid coordinates based on the geographic coordinates of all interrupted regular supply points under each original regular supply point s interruption scenario. And the average discrete distance of each interrupted regular supply point relative to the spatial centroid coordinates. .

[0153] This reflects the degree of spatial clustering or dispersion of points disrupting regular supply. The larger the average dispersion distance, the more geographically dispersed the points disrupting regular supply; the smaller the average dispersion distance, the more concentrated the points disrupting regular supply.

[0154] Step S703: Combine the comprehensive loss score, spatial centroid coordinates, and average discrete distance to form a space-capacity joint feature vector.

[0155] In one possible implementation, the space-capacity joint eigenvector is represented as:

[0156] Among them, the comprehensive loss score Reflects the degree of supply capacity loss (capacity dimension), spatial centroid coordinates and average discrete distance It reflects the spatial distribution characteristics (spatial dimension) of the points where regular supply is interrupted. The four components together constitute the feature vector used for clustering reduction.

[0157] According to this embodiment, by constructing a spatial-capacity joint feature vector that includes the comprehensive loss score, spatial centroid coordinates, and average discrete distance, the clustering reduction process can simultaneously retain information on two dimensions: the degree of supply capacity loss and the spatial distribution of interruption points, thus ensuring that the representative scene set fully characterizes the features of the original scene.

[0158] According to one embodiment of this application, an emergency supply chain network design system is provided for performing the methods described in the above embodiments.

[0159] like Figure 3 As shown, the emergency supply chain network design system includes a data acquisition module 301, a network construction module 302, a scenario generation and reduction module 303, and an optimization solution module 304. The system reads parameters from emergency resource databases, geographic information databases, disaster demand databases, and historical disaster data, and outputs emergency material reserve layout, cooperative relationships, inventory configuration, and allocation schemes under different interruption scenarios after model calculation.

[0160] The data acquisition module 301 is used to acquire parameter data of each node within the disaster area. The nodes include regular supply points, temporary distribution centers, disaster-stricken areas, and backup supply points. The parameter data includes: the location of each node, the supply capacity or storage capacity of various emergency supplies, the transportation distance and time between nodes, and the emergency supply demand of disaster-stricken areas.

[0161] In one possible implementation, the data acquisition module 301 connects to an electronic map platform, converts node location data into latitude and longitude coordinates, and calculates the transportation distance and time between nodes based on the coordinates; it also connects to a disaster demand database to calculate the demand for various emergency supplies based on the size of the affected population and the per capita material demand coefficient.

[0162] Network building module 302 is used to build an initial emergency supply chain network based on parameter data. The initial emergency supply chain network includes regular supply logistics connections from regular supply points to temporary distribution centers, distribution logistics connections from temporary distribution centers to disaster-stricken areas, and direct supplementary distribution logistics connections from backup supply points to disaster-stricken areas.

[0163] In one possible implementation, network building module 302 constructs an initial network according to a resilience strategy: each temporary distribution center establishes regular supply logistics connections with at least two regular supply points when it is activated; safety stock of various emergency supplies is pre-positioned in each temporary distribution center; and backup supply points establish cooperative relationships with disaster-stricken areas before the disaster.

[0164] The scenario generation and reduction module 303 is used to generate all original regular supply point interruption scenarios based on the number of interruptions at regular supply points, and to cluster and reduce all original regular supply point interruption scenarios to obtain a representative set of regular supply point interruption scenarios.

[0165] In one possible implementation, the scene generation and reduction module 303 first divides the interruption scenarios into different interruption orders based on the number of regular supply point interruptions, and generates all original regular supply point interruption scenarios under each interruption order by enumeration. Then, it constructs a space-capacity joint feature vector for each original regular supply point interruption scenario, and uses the k-medoids clustering algorithm to cluster and reduce all original scenarios. The original scenarios belonging to the same representative scenario are merged to obtain a set of representative regular supply point interruption scenarios and their occurrence probabilities.

[0166] The optimization solution module 304 is used to optimize the pre-disaster resource allocation decision and the emergency material allocation decision under the post-disaster representative conventional supply point interruption scenarios based on the set of representative conventional supply point interruption scenarios and the initial emergency supply chain network, so as to obtain the emergency supply chain network design scheme corresponding to the disaster area.

[0167] In one possible implementation, the optimization solution module 304 achieves optimization by constructing and solving a bi-objective, two-stage stochastic mixed-integer programming model, as detailed in Examples 2 to 4. The optimization solution module 304 ultimately outputs the optimal values ​​for each decision variable, forming a complete emergency supply chain network design scheme.

[0168] According to this embodiment, by setting up functional modules corresponding to each method step, the emergency supply chain network design method is systematically integrated, which can automatically complete the entire process from data acquisition, network construction, scenario generation to optimization solution, thereby improving the efficiency and accuracy of emergency supply chain network design; the modules communicate with each other through standardized data interfaces, which facilitates system maintenance and upgrades.

[0169] According to one embodiment of this application, an electronic device is provided for performing the methods described in the above embodiments. Figure 4This is a schematic diagram of an electronic device according to an embodiment of the present application. The electronic device includes a processor 110, a communication interface 120, a memory 130, and a communication bus 140. The processor 110, the communication interface 120, and the memory 130 communicate with each other via the communication bus 140.

[0170] The communication bus 140 is used to realize data communication between the processor 110, the communication interface 120, and the memory 130. The communication bus 140 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect Standard (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus 140 can be divided into an address bus, a data bus, a control bus, etc.

[0171] Communication interface 120 is used to enable data communication between electronic devices and external devices. In one possible implementation, communication interface 120 connects to one or more of the following external devices: an emergency resource database (providing data such as the location of regular supply points, reserve capacity, and inter-node transportation parameters), a geographic information database (providing latitude and longitude coordinates and spatial analysis data), a disaster demand database (providing data on the location of disaster-stricken points and demand), and a historical disaster database (providing data such as the probability distribution of interruption scenarios). In another possible implementation, communication interface 120 is also used to receive user-input decision preference information (including a weighted combination of economic costs and comprehensive deprivation costs) to determine the final emergency supply chain network design scheme.

[0172] In one possible implementation, the communication interface 120 can be a wired communication interface (such as an Ethernet interface) or a wireless communication interface (such as a WiFi interface, Bluetooth interface, mobile communication interface, etc.).

[0173] The memory 130 is used to store computer programs and related data. In one possible implementation, the memory 130 stores one or more of the following data: parameter data of each node in the disaster area (including the location of each node, the supply capacity or storage capacity of various emergency supplies, the transportation distance and transportation time between each node, and the emergency supply demand of the disaster-stricken point); data on regular supply point interruption scenarios (including the original set of regular supply point interruption scenarios and the representative set of regular supply point interruption scenarios obtained after clustering reduction and their probability of occurrence); model parameter data (including economic cost parameters and comprehensive deprivation cost parameters); candidate solutions, Pareto non-dominated solution sets and optimal solutions of decision variables generated during the solution process; and the final emergency supply chain network design scheme.

[0174] In one possible implementation, memory 130 is any combination of one or more of the following types: random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, hard disk drive (HDD), solid-state drive (SSD), or cloud storage. Memory 130 may exist independently of processor 110 or may be connected to processor 110 via communication bus 140. Memory 130 may also be integrated with processor 110.

[0175] The processor 110 is used to execute the computer program stored in the memory 130 to implement the steps of the emergency supply chain network design method described in the embodiments of this application.

[0176] In one possible implementation, the processor 110 is any of the following types: central processing unit (CPU), graphics processing unit (GPU), digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other programmable logic device.

[0177] In one possible implementation, the electronic device is any one of a server, computer, cloud computing platform, emergency management decision support system terminal, or logistics dispatching system terminal. The electronic device has data processing and optimization computing capabilities and is able to run the emergency supply chain network design method described in this application.

[0178] In practical use, the system reads parameters from the emergency resource database, geographic information database, disaster demand database, and historical disaster database through the communication interface 120. After the processor 110 executes the computer program stored in the memory 130 to perform model calculations, it outputs the emergency material reserve layout, cooperation relationships, inventory configuration, and allocation schemes under different interruption scenarios through the communication interface 120. The output results can be displayed on a display device or stored in the memory 130 for later review.

[0179] According to this embodiment, by storing the emergency supply chain network design method described in this application as a computer program in the memory of an electronic device and executing it through a processor, the method is automated, improving the efficiency and accuracy of emergency supply chain network design. The electronic device has data processing and optimization computing capabilities, enabling it to process large amounts of node data, data on regular supply point interruption scenarios, and model parameter data in a short time, quickly generating the optimal emergency supply chain network design scheme. By connecting to various external databases and external devices through communication interfaces, the system automates data input, centralizes data storage, and makes the scheme output more intuitive, facilitating its use by emergency management personnel in practical work.

[0180] An exemplary embodiment of this application is described below using the area surrounding the Longmenshan Fault Zone in Sichuan Province as an example. The Longmenshan Fault Zone is located in a region of frequent seismic activity in my country, and has experienced numerous destructive earthquakes. The capacity for emergency material support directly impacts rescue efficiency and the living standards of affected populations after a disaster. Therefore, this embodiment selects some urban nodes surrounding the Longmenshan Fault Zone as the research object, constructing a multi-level emergency supply chain network consisting of regular supply points, temporary distribution centers, disaster-stricken areas, and backup supply points.

[0181] In the data preparation phase, the geographical location, reserve capacity, and demand information of regular supply points, candidate temporary distribution centers, disaster-stricken areas, and backup supply points were first acquired. The relevant basic node data are shown in Tables 1 to 4 below, where Table 1 contains data related to disaster-stricken areas, Table 2 contains data related to regular supply points, Table 3 contains data related to temporary distribution centers, and Table 4 contains data related to backup supply points. The geographical location data for each node was obtained from an electronic map platform and uniformly converted to latitude and longitude coordinates. The relevant cost and demand parameters were mainly based on existing research and appropriately adjusted to suit the scale of this example. The transportation distances between regular supply points and temporary distribution centers, between temporary distribution centers and disaster-stricken areas, and between backup supply points and disaster-stricken areas were calculated based on the node coordinates, and the transportation time parameters were calculated using the average vehicle speed. In this embodiment, the average vehicle speed was uniformly set to 70 km / h.

[0182]

[0183] Table 1

[0184] Table 2

[0185] Table 3

[0186] Table 4 This embodiment considers three typical emergency supplies: drinking water, food, and tents. Table 5 shows the pre-disaster procurement unit prices for each type of supply. Post-disaster procurement unit prices from regular supply points Price of supplies procured from backup supply points after a disaster Unit inventory holding costs at temporary distribution centers Unit distance transportation cost Deprivation cost function parameters and The demand for various supplies by the population in the disaster-stricken area and minimum demand satisfaction rate The safety stock ratio is set at 0.2 to determine the pre-positioned inventory size of temporary distribution centers; the relative takeaway cost weighting coefficient in the comprehensive takeaway cost model is set at 0.75 to balance the impact of absolute takeaway cost and relative takeaway cost on the overall takeaway cost.

[0187]

[0188] Table 5 Regarding network construction costs, the fixed activation cost for both regular supply points and temporary distribution centers is set at 200,000 yuan, while the fixed activation cost for backup supply points is set at 500,000 yuan. Path cooperation agreement costs are set according to different cooperation relationships: the cooperation agreement cost between a regular supply point and a temporary distribution center is 50,000 yuan; the cooperation agreement cost between a temporary distribution center and a disaster-stricken area is 40,000 yuan; the cooperation agreement cost between a backup supply point and a disaster-stricken area is 80,000 yuan; and the safety stock contract cost is set at 150,000 yuan. These parameters are used to characterize the investment required for pre-disaster network construction, establishing cooperation relationships, and pre-positioning inventory, providing basic data support for subsequent economic cost analysis.

[0189] To simulate the impact of disasters on the supply system, this embodiment constructs different levels of regular supply point interruption scenarios based on the number of interruptions at these points, including scenarios with no regular supply point interruptions, one regular supply point interruption, two regular supply point interruptions, and three regular supply point interruptions. The occurrence probabilities corresponding to each interruption order are set to 0.45, 0.35, 0.15, and 0.05, respectively, to reflect the disaster risk characteristic of higher probability of low-order interruptions and lower probability of high-order interruptions. Based on this, an enumeration method is used to generate all possible combinations of regular supply point interruptions, resulting in 93 original interruption scenarios. For multiple original scenarios generated under the same interruption order, the total occurrence probability corresponding to that interruption order is evenly distributed among the original scenarios to obtain the occurrence probability for each scenario.

[0190] Due to the large number of original scenarios, directly incorporating them into the stochastic programming model would significantly increase the solution scale. Therefore, a scenario reduction method based on joint spatial-capacity features is further adopted. First, the spatial distribution features and supply capacity loss features corresponding to each scenario are extracted, and scenario feature vectors are constructed. Then, the feature vectors are standardized, and the k-medoids clustering method is used to reduce the scenarios within the same interruption order. In this embodiment, the clustering number parameter k is set to 3, that is, 3 representative scenarios are retained for each interruption order. After clustering reduction, 10 representative regular supply point interruption scenarios are finally retained, and their specific information is shown in Table 6 below. As can be seen from Table 6, the reduced scenario set can better retain the main features of the original scenarios in terms of interruption order, spatial distribution, and supply capacity loss, while significantly reducing the solution scale of the stochastic programming model.

[0191]

[0192] Table 6 In the model solution phase, a two-stage stochastic mixed integer programming model with the objectives of minimizing economic cost and comprehensive deprivation cost was established based on the above data. The model was then optimized using the Gurobi solver within a Python programming environment. Since there is a clear trade-off between the economic cost objective and the comprehensive deprivation cost objective, an adaptive approach was adopted. The constraint method is used to construct the Pareto non-dominated solution set. First, the economic cost minimization model and the comprehensive deprivation cost minimization model are solved separately to obtain two anchor solutions, and the range of the comprehensive deprivation cost objective is determined accordingly. Then, the comprehensive deprivation cost objective is transformed into... The constraints are used to solve the problem with minimizing economic cost as the primary optimization objective, and the initial values ​​are uniformly set within the specified range. Find the value point and obtain the initial candidate solution set.

[0193] Further non-dominated screening is performed on the candidate solution set, and dominated solutions are removed to form the initial Pareto non-dominated solution set. Then, the target value distance between adjacent non-dominated solutions is calculated in the economic cost dimension and the comprehensive deprivation cost dimension, and the results are adjusted based on the distance. An adaptive encryption solution is applied to the intervals to obtain the Pareto non-dominated solution set, as shown in the following figure. Figure 5 As shown. Figure 5 Each point in the equation corresponds to a Pareto nondominated solution, reflecting the trade-off between economic costs and comprehensive deprivation costs. Figure 5 It can be seen that as economic costs increase, the overall cost of deprivation shows a downward trend, indicating that improving disaster protection levels usually requires higher network construction and operation costs.

[0194] To further analyze the characteristics of different decision preferences, this embodiment selects five representative schemes from the Pareto frontier: the lowest economic cost scheme, the partial economic cost scheme, the equilibrium compromise scheme, the partial deprivation cost scheme, and the lowest deprivation cost scheme. In the selection process, the economic cost and the overall deprivation cost are first normalized, and then weights are set according to different decision preferences, and the overall evaluation value is calculated. The lowest economic cost scheme and the lowest deprivation cost scheme are directly determined by the optimal solution of their respective objective functions; the partial economic cost scheme, the equilibrium compromise scheme, and the partial deprivation cost scheme are determined by the non-dominated solution with the optimal overall evaluation value under different weight combinations. The network design results and objective function values ​​corresponding to each representative scheme are shown in Table 7.

[0195] As shown in Table 7, the network design schemes obtained under different decision preferences exhibit significant differences in the activation of regular supply points, construction of temporary distribution centers, configuration of safety stock, and mobilization of backup supply points. Among these, the lowest economic cost scheme minimizes network construction and operation costs, but has a higher overall deprivation cost; the lowest deprivation cost scheme significantly improves disaster preparedness, but requires more construction investment; and the equilibrium compromise scheme achieves a good balance between economic cost and disaster relief capacity, demonstrating the effectiveness and applicability of the proposed method in multi-objective emergency supply chain network design.

[0196]

[0197] Table 7

Claims

1. An emergency supply chain network design method, characterized in that, Includes the following steps: Acquire parameter data for each node within the disaster area, including regular supply points, temporary distribution centers, disaster-stricken areas, and backup supply points; the parameter data includes: the location of each node, the supply capacity or storage capacity of various emergency supplies, the transportation distance and time between each node, and the emergency supply demand of the disaster-stricken areas; An initial emergency supply chain network is constructed based on the parameter data. The initial emergency supply chain network includes regular supply logistics connections from regular supply points to temporary distribution centers, distribution logistics connections from temporary distribution centers to disaster-stricken areas, and direct supplementary distribution logistics connections from backup supply points to disaster-stricken areas. Based on the number of interruptions at regular supply points, all original regular supply point interruption scenarios are generated, and all the original regular supply point interruption scenarios are clustered and reduced to obtain a representative set of regular supply point interruption scenarios. Based on the set of representative regular supply point interruption scenarios and the initial emergency supply chain network, the pre-disaster resource allocation decisions and the post-disaster emergency material allocation decisions under the representative regular supply point interruption scenarios are optimized to obtain the emergency supply chain network design scheme corresponding to the disaster area.

2. The method according to claim 1, characterized in that, Based on the set of representative conventional supply point disruption scenarios and the initial emergency supply chain network, the pre-disaster resource allocation decisions and post-disaster emergency material allocation decisions under various representative conventional supply point disruption scenarios are optimized to obtain the emergency supply chain network design scheme corresponding to the disaster area, including: Based on the representative set of regular supply point disruption scenarios and the initial emergency supply chain network, a bi-objective two-stage stochastic mixed integer programming model is solved to obtain the optimal solution for the decision variables. The bi-objective two-stage stochastic mixed integer programming model has a first-stage variable characterizing pre-disaster resource allocation decisions and a second-stage variable characterizing emergency material allocation decisions under post-disaster disruption scenarios of various representative regular supply points. The bi-objective two-stage stochastic mixed integer programming model has two optimization objectives: minimizing economic costs and minimizing comprehensive deprivation costs. Among them, the pre-disaster resource allocation decisions include the activation decisions of regular supply points, temporary distribution centers, and backup supply points; the establishment decisions of cooperative relationships between regular supply points and temporary distribution centers, between temporary distribution centers and disaster-stricken areas, and between backup supply points and disaster-stricken areas; and the decisions on the safety stock allocation of various emergency materials within temporary distribution centers. Based on the optimal solution of the decision variables, the design scheme of the emergency supply chain network is determined.

3. The method according to claim 2, characterized in that, Based on the representative set of regular supply point disruption scenarios and the initial emergency supply chain network, a two-stage stochastic mixed-integer programming model with dual objectives is solved to obtain the optimal solution for the decision variables, including: S1. Solve the model with the single objectives of minimizing economic cost and minimizing comprehensive deprivation cost respectively to obtain two anchor solutions, and determine the range of values ​​for the comprehensive deprivation cost objective based on the two anchor solutions. S2. Set multiple ε value points within the range of values, transform the comprehensive deprivation cost objective into an ε constraint, and solve the problem with minimizing economic cost as the main objective at each ε value point to obtain the corresponding candidate solution; S3. Perform non-dominated screening on all the candidate solutions, delete dominated solutions, and form an initial Pareto non-dominated solution set; S4. Based on the spacing distribution of adjacent solutions in the target space in the initial Pareto non-dominated solution set, identify sparse target space regions, insert new ε value points in the ε interval corresponding to the target space region, re-solve the corresponding ε-constrained subproblem, and add the obtained new solution to the candidate solution. S5. Perform non-dominated screening on the candidate solutions after adding the new solutions, and update the Pareto non-dominated solution set; S6. Repeat steps S4 to S5 until the preset termination condition is met, and obtain the target Pareto non-dominated solution set as the optimal solution of the decision variable.

4. The method according to claim 3, characterized in that, The process of determining the emergency supply chain network design scheme based on the optimal solution of decision variables includes: The economic cost and total deprivation cost corresponding to each non-dominated solution in the target Pareto non-dominated solution set are normalized to obtain the normalized economic cost and normalized total deprivation cost corresponding to each non-dominated solution. Obtain the decision weight combination corresponding to various decision preferences, including the lowest economic cost preference, the partial economic cost preference, the equilibrium compromise preference, the partial deprivation cost preference, and the lowest deprivation cost preference; For each decision preference, based on the corresponding decision weight combination, the normalized economic cost and the normalized comprehensive deprivation cost corresponding to each non-dominated solution, the comprehensive evaluation value of each non-dominated solution corresponding to the decision preference is calculated. For each decision preference, the non-dominated solution with the smallest comprehensive evaluation value among all non-dominated solutions corresponding to the decision preference is determined as the optimal solution corresponding to the decision preference; Based on the optimal solutions corresponding to various decision preferences, the emergency supply chain network design scheme is determined; the emergency supply chain network design scheme includes one or more of the following: minimum economic cost solution, partial economic cost solution, equilibrium compromise solution, partial deprivation cost solution, and minimum deprivation cost solution.

5. The method according to claim 2, characterized in that, Based on the aforementioned parameter data, an initial emergency supply chain network is constructed, including: Based on the resilience strategy combination and the parameter data, the initial emergency supply chain network is constructed. The resilience strategy combination includes: when the temporary distribution center is activated, it establishes regular supply logistics connections with at least two regular supply points; various types of emergency supplies are pre-positioned in safety stock at each of the temporary distribution centers, and the configuration amount of the safety stock is determined by the safety stock configuration amount in the first stage variables; the backup supply points establish cooperative relationships with disaster-stricken areas before the disaster and are triggered to be activated when the regular supply capacity is insufficient after the disaster, directly delivering emergency supplies to the disaster-stricken areas.

6. The method according to claim 1, characterized in that, The representative set of regular supply point interruption scenarios includes at least one of the representative regular supply point interruption scenarios; the clustering and reduction of all the original regular supply point interruption scenarios to obtain the representative set of regular supply point interruption scenarios includes: For each original regular supply point interruption scenario, a feature vector is constructed to characterize that original regular supply point interruption scenario; The feature vectors of each original regular supply point interruption scenario are standardized, and a clustering algorithm is used to divide all original regular supply point interruption scenarios into several classes; The original regular supply point interruption scenarios corresponding to the cluster centers of each type are selected as candidate regular supply point interruption scenarios; The original regular supply point interruption scenarios belonging to the same candidate regular supply point interruption scenario are merged to obtain at least one representative regular supply point interruption scenario.

7. The method according to claim 6, characterized in that, For each original regular supply point interruption scenario, a feature vector is constructed to characterize that original regular supply point interruption scenario, including: A comprehensive loss score is calculated based on the loss of supply capacity of various emergency supplies under each original regular supply point interruption scenario and the preset weight of material importance. Based on the geographical coordinates of all interrupted regular supply points under each original regular supply point interruption scenario, calculate the spatial centroid coordinates and the average discrete distance of each interrupted regular supply point relative to the spatial centroid coordinates. The combined loss score, spatial centroid coordinates, and average discrete distance are combined to form the space-capacity joint feature vector.

8. The method according to claim 2, characterized in that, The economic costs include pre-disaster costs and expected post-disaster costs; The pre-disaster costs include the fixed activation costs of regular supply points, the fixed activation costs of temporary distribution centers, the fixed activation costs of backup supply points, the costs of establishing cooperative relationships between nodes, the costs of safety stock contracts, the costs of safety stock procurement, the costs of safety stock transportation, and the costs of holding safety stock. The expected post-disaster costs include the procurement costs of emergency supplies through conventional and backup supply channels under different representative scenarios of disruption at conventional supply points, as well as transportation costs during conventional supply, safety stock mobilization, and backup supply processes; and / or, The comprehensive deprivation cost includes the absolute deprivation cost at the disaster-stricken site and the relative deprivation cost between disaster-stricken sites due to differences in protection levels; wherein, The absolute deprivation cost is calculated using an exponential unit deprivation cost function based on the unmet demand at the disaster site and the maximum scarcity time; the maximum scarcity time is determined based on the maximum transport time of all feasible transport routes in the emergency supply chain network.

9. An emergency supply chain network design system, characterized in that, include: The data acquisition module is used to acquire parameter data of each node in the disaster area. The nodes include regular supply points, temporary distribution centers, disaster-stricken areas, and backup supply points. The parameter data includes: the location of each node, the supply capacity or storage capacity of various emergency supplies, the transportation distance and transportation time between each node, and the emergency supply demand of the disaster-stricken areas. The network construction module is used to construct an initial emergency supply chain network based on the parameter data. The initial emergency supply chain network includes regular supply logistics connections from regular supply points to temporary distribution centers, distribution logistics connections from temporary distribution centers to disaster-stricken areas, and direct supplementary distribution logistics connections from backup supply points to disaster-stricken areas. The scenario generation and reduction module is used to generate all original regular supply point interruption scenarios based on the number of interruptions at regular supply points, and to cluster and reduce all the original regular supply point interruption scenarios to obtain a representative set of regular supply point interruption scenarios. The optimization solution module is used to optimize the pre-disaster resource allocation decision and the post-disaster emergency material allocation decision under the representative set of interruption scenarios of regular supply points and the initial emergency supply chain network, so as to obtain the emergency supply chain network design scheme corresponding to the disaster area.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable by the processor, characterized in that, The processor implements the method as described in any one of claims 1 to 8 when executing the computer program.

Images