Emergency resource scheduling optimization simulation method and system for global disaster situations

Abstract

The invention relates to the technical field of resource scheduling, and particularly discloses an emergency resource scheduling optimization simulation method and system for a global disaster situation, and the method comprises the steps: integrating disaster environment data, emergency resource data, traffic network data and historical disaster situation data, and defining and parameterizing a global disaster situation scene; establishing an emergency resource scheduling optimization model containing a spatio-temporal evolution process, and generating at least one preliminary scheduling scheme by using an optimization algorithm; the generated preliminary scheduling scheme is placed in a simulation environment based on multiple agents and a geographic information system to execute a deduction process, a comprehensive evaluation index is obtained through calculation, and when the comprehensive evaluation index is lower than a preset satisfaction threshold, a feedback mechanism is triggered; recursively optimizing the emergency resource scheduling optimization model and circularly executing, and outputting a final scheduling scheme; according to the method, comprehensive and objective quantitative evaluation is carried out on the preliminary scheduling scheme by integrating multi-dimensional indexes, so that a resource scheduling scheme which better meets actual requirements is generated in a recursive manner.

Description

Technical Field

[0001] This invention relates to the field of resource scheduling technology, specifically to an emergency resource scheduling optimization simulation method and system for global disaster situations. Background Technology

[0002] In recent years, natural disasters and public safety emergencies have occurred frequently around the world, posing unprecedented challenges to the ability to dispatch emergency resources. Emergency resource dispatch systems must coordinate multiple resources, multi-stage tasks, and multi-objective constraints in highly complex, dynamically evolving, and uncertain disaster environments in order to maximize the effectiveness of rescue efforts.

[0003] However, traditional scheduling methods generally rely on static mathematical models or decision-makers' experience and judgment, which are difficult to effectively adapt to the rapid evolution of disasters in the spatiotemporal dimensions, the real-time dynamic damage of transportation networks, and the drastic fluctuations in resource demand, among other uncertainties. In terms of evaluation systems, existing methods usually use a single or a few indicators for evaluation, failing to systematically integrate multi-dimensional performance such as the timeliness of resource delivery, the sufficiency of demand satisfaction, the fairness of spatial allocation, the robustness of the plan in dealing with uncertainties, and economic cost efficiency. This results in one-sided evaluation results that cannot fully reflect the comprehensive effectiveness of the scheduling plan. Therefore, how to construct a comprehensive and accurate resource scheduling plan rapid generation technology is the technical problem that this invention aims to solve. Summary of the Invention

[0004] The purpose of this invention is to provide a simulation method and system for optimizing emergency resource scheduling in a global disaster situation, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A simulation method and system for optimizing emergency resource scheduling in a comprehensive disaster situation, the method comprising: S1: Integrate disaster environment data, emergency resource data, transportation network data, and historical disaster data to define and parameterize the overall disaster scenario; S2: Based on the scenario constructed in step S1, establish an emergency resource scheduling optimization model containing spatiotemporal evolution process, and use optimization algorithms to generate at least one preliminary scheduling scheme. S3: Place the preliminary scheduling scheme generated in step S2 into a simulation environment based on multi-agent and geographic information system to perform the simulation process, simulating resource flow, traffic congestion, dynamic changes in demand and unexpected events. S4: Construct a multi-dimensional comprehensive evaluation model, process the raw data output from the simulation in step S3, calculate the comprehensive evaluation index, and trigger the feedback mechanism when the comprehensive evaluation index is lower than the preset satisfaction threshold. S5: Recursively optimize the emergency resource scheduling optimization model in step S2 based on the evaluation results of step S4; the optimization methods include adjusting the objective function weights and modifying the model parameters; S6: Repeat steps S2 to S5 until the preset loop exit condition is met, and output the final scheduling scheme and a complete simulation evaluation report; the loop exit condition includes the comprehensive evaluation index reaching the satisfaction threshold and the number of loops reaching the preset maximum number of iterations.

[0006] As a further aspect of the present invention: the disaster environment data includes geospatial data and real-time disaster evolution prediction data; the emergency resource data includes the resource reserves, types, and locations of multi-level reserve warehouses, as well as the attributes and quantity of transport vehicles; the transportation network data includes the road network topology, traffic capacity, and dynamic damage probability.

[0007] As a further aspect of the present invention: the emergency resource scheduling optimization model includes a multi-stage stochastic programming model and a robust optimization model, and its objective function includes at least minimizing the total response time, maximizing the fairness of demand satisfaction, and minimizing the total system cost, and its decision variables include resource allocation, transportation routes, and inventory dynamics.

[0008] As a further aspect of the present invention: the simulation environment based on multi-agent and geographic information systems in step S3 includes: The environment layer is a static geospatial and transportation network model based on GIS. The entity layer includes intelligent agents at disaster sites, intelligent agents for transport vehicles, intelligent agents for resources, and intelligent agents for the command center. Each intelligent agent has autonomous behavior rules and interaction logic. The process layer simulates the dynamic processes of resource loading and unloading, transportation navigation, traffic flow, and information transmission delays.

[0009] As a further aspect of the present invention: the process of calculating the comprehensive evaluation index in step S4 is as follows: The efficiency index, effectiveness index, fairness index, economic index, and economic index are weighted and summed to obtain the comprehensive evaluation index. The satisfaction threshold is connected to a dynamic adjustment port, which is used to adjust according to the disaster level and rescue stage.

[0010] As a further aspect of the present invention: the efficiency index is obtained as follows: The actual delivery time of resources to all disaster-stricken areas was statistically analyzed and compared with the expected delivery time for each area. Calculate the delay time at each point and normalize it based on the maximum allowable response time of the system. The overall efficiency index is obtained by averaging the timeliness scores of all disaster-affected locations. The effectiveness index is obtained as follows: The calculation requires summarizing the actual delivery quantities of all disaster-stricken areas and all types of resources, and comparing them with the actual demand of the corresponding areas. The smaller of the actual delivery quantity and the demand quantity is summed and then divided by the total demand quantity to obtain the effectiveness index. The fairness index is obtained as follows: The calculation is based on the distribution of resource satisfaction rate at each disaster site, and the Gini coefficient or Theil index is used to measure the degree of imbalance. The fairness index is determined based on the inverse ratio of the degree of imbalance. The robustness index is obtained as follows: In the simulation, a preset number of disturbance scenarios are simulated, the same scheduling scheme is executed respectively, and their overall performance is recorded; The robustness index is calculated by comparing the evaluation results under each disturbance scenario with the undisturbed baseline scenario. The economic indices are obtained as follows: The total cost of implementing the statistical scheme, including transportation expenses, warehousing expenses, labor expenses and management expenses, is compared with the upper and lower limits of costs in the simulation environment, and the economic index is obtained through linear normalization.

[0011] The present invention also provides a simulation system for optimizing emergency resource scheduling in a global disaster situation, the system comprising: The scenario parameterization module is used to integrate disaster environment data, emergency resource data, transportation network data, and historical disaster data to define and parameterize disaster scenarios across the entire domain. The scheduling scheme generation module is used to establish an emergency resource scheduling optimization model containing spatiotemporal evolution process based on the scenario constructed by the scenario parameterization module, and to generate at least one preliminary scheduling scheme using optimization algorithms. The scheme simulation module is used to put the preliminary scheduling scheme generated by the scheduling scheme generation module into a simulation environment based on multi-agent and geographic information system to perform the simulation process, simulating resource flow, traffic congestion, dynamic changes in demand and unexpected events. The data analysis module is used to construct a multi-dimensional comprehensive evaluation model, process the raw data output from the scheme deduction module simulation, calculate the comprehensive evaluation index, and trigger a feedback mechanism when the comprehensive evaluation index is lower than the preset satisfaction threshold. The model optimization module is used to recursively optimize the emergency resource scheduling optimization model in the scheduling scheme generation module based on the evaluation results of the data analysis module; the optimization methods include adjusting the objective function weights and modifying the model parameters. The loop execution module is used to repeatedly execute the scenario parameterization module to the model optimization module until the preset loop exit conditions are met, and output the final scheduling scheme and a complete simulation evaluation report; the loop exit conditions include the comprehensive evaluation index reaching the satisfaction threshold and the number of loops reaching the preset maximum number of iterations.

[0012] As a further aspect of the present invention: the disaster environment data includes geospatial data and real-time disaster evolution prediction data; the emergency resource data includes the resource reserves, types, and locations of multi-level reserve warehouses, as well as the attributes and quantity of transport vehicles; the transportation network data includes the road network topology, traffic capacity, and dynamic damage probability.

[0013] As a further aspect of the present invention: the emergency resource scheduling optimization model includes a multi-stage stochastic programming model and a robust optimization model, and its objective function includes at least minimizing the total response time, maximizing the fairness of demand satisfaction, and minimizing the total system cost, and its decision variables include resource allocation, transportation routes, and inventory dynamics.

[0014] As a further aspect of the present invention: the simulation environment based on multi-agent and geographic information systems in the scheme deduction module includes: The environment layer is a static geospatial and transportation network model based on GIS. The entity layer includes intelligent agents at disaster sites, intelligent agents for transport vehicles, intelligent agents for resources, and intelligent agents for the command center. Each intelligent agent has autonomous behavior rules and interaction logic. The process layer simulates the dynamic processes of resource loading and unloading, transportation navigation, traffic flow, and information transmission delays.

[0015] Compared with the prior art, the beneficial effects of the present invention are: This invention provides a systematic and multi-dimensional evaluation framework that solves the problems of existing methods' one-sided evaluation and inability to fully reflect the comprehensive effectiveness of dispatching schemes. By integrating multiple key indicators such as efficiency, effectiveness, fairness, robustness, and economy, it can conduct a comprehensive and objective quantitative evaluation of preliminary dispatching schemes, avoiding the limitations of single-indicator evaluation. At the same time, by setting adjustable weights for each dimension index, the evaluation model can flexibly adapt to different disaster levels, rescue stages, and decision preferences, thereby generating a comprehensive evaluation index that better meets actual needs. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention.

[0017] Figure 1 The overall flowchart of the simulation method for optimizing emergency resource scheduling in a global disaster situation is shown.

[0018] Figure 2 The diagram shows the structure of a simulation system for optimizing emergency resource allocation in a global disaster situation. Detailed Implementation

[0019] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0020] Figure 1 The present invention provides a flowchart of an emergency resource scheduling optimization simulation method and system for comprehensive disaster situations. In this embodiment, an emergency resource scheduling optimization simulation method for comprehensive disaster situations includes: S1: Integrate disaster environment data, emergency resource data, transportation network data, and historical disaster data to define and parameterize the overall disaster scenario; A comprehensive disaster scenario refers to a complete, dynamic, and parameterized description of the occurrence, development, and impact of a disaster by integrating multi-source data. It encompasses elements such as the geographical scope of the disaster, its temporal evolution, the severity of the disaster, the distribution of resource demands, and the condition of the transportation network, providing fundamental data support for subsequent scheduling optimization and simulation.

[0021] S2: Based on the scenario constructed in step S1, establish an emergency resource scheduling optimization model containing spatiotemporal evolution process, and use optimization algorithms to generate at least one preliminary scheduling scheme. The emergency resource scheduling optimization model is a multi-stage, multi-objective emergency resource scheduling optimization model. It is a mathematical programming model designed to solve the emergency resource allocation and transportation problem at different time stages, simultaneously optimizing multiple conflicting objectives (such as minimizing response time, maximizing demand satisfaction rate, and minimizing cost). This model can capture the dynamic evolution of disaster situations and generate highly adaptive scheduling schemes.

[0022] S3: Place the preliminary scheduling scheme generated in step S2 into a simulation environment based on multi-agent and geographic information system to perform the simulation process, simulating resource flow, traffic congestion, dynamic changes in demand and unexpected events. A simulation environment based on multi-agent systems and geographic information systems refers to a virtual, high-fidelity simulation platform. Geographic Information Systems (GIS) provide a realistic geographic space and transportation network background, while the multi-agent system simulates the autonomous behavior and interactions of different entities such as disaster sites, transport vehicles, resources, and command centers, dynamically extrapolating the execution process of scheduling plans to reflect resource flow, traffic congestion, demand changes, and the impact of emergencies.

[0023] S4: Construct a multi-dimensional comprehensive evaluation model, process the raw data output from the simulation in step S3, calculate the comprehensive evaluation index, and trigger the feedback mechanism when the comprehensive evaluation index is lower than the preset satisfaction threshold. A multi-dimensional comprehensive evaluation model is an evaluation framework used to quantify the overall performance of scheduling schemes. This model integrates multiple key performance indicators (such as efficiency, effectiveness, fairness, robustness, and economy) to systematically analyze simulation results, providing a comprehensive and objective evaluation of the schemes.

[0024] The comprehensive evaluation index is a single numerical value calculated by a multi-dimensional comprehensive evaluation model, used to quantify the overall merits of emergency resource allocation plans. This index comprehensively reflects the plan's performance across multiple dimensions, including efficiency, effectiveness, fairness, robustness, and economy.

[0025] The satisfaction threshold refers to the minimum acceptable level of a pre-set comprehensive evaluation index. When the comprehensive evaluation index of a dispatch plan reaches or exceeds this threshold, the plan is considered to meet the expected performance requirements. This threshold can be dynamically adjusted according to the specific disaster level and rescue stage.

[0026] S5: Recursively optimize the emergency resource scheduling optimization model in step S2 based on the evaluation results of step S4; the optimization methods include adjusting the objective function weights and modifying the model parameters; In the above process, the feedback mechanism refers to the adjustment process that is automatically or manually triggered by the system when the evaluation result of the scheduling scheme does not meet the satisfaction threshold. This mechanism guides the optimization algorithm to generate a better scheduling scheme by modifying the objective function weights or model parameters of the optimization model, thereby realizing dynamic correction and continuous improvement of the scheduling process.

[0027] S6: Repeat steps S2 to S5 until the preset loop exit condition is met, and output the final scheduling scheme and a complete simulation evaluation report; the loop exit condition includes the comprehensive evaluation index reaching the satisfaction threshold and the number of loops reaching the preset maximum number of iterations.

[0028] In one example of the technical solution of this invention, by integrating multi-source disaster data and parameterizing scenarios, a digital solution is provided for the scheduling optimization process. By establishing a multi-stage, multi-objective optimization model that considers spatiotemporal evolution, a preliminary scheduling scheme is generated, improving the adaptability of the scheme. The scheme is then placed in a simulation environment based on multi-agent and geographic information systems for deduction, which can simulate complex dynamic environments and verify the execution effect of the scheme. A comprehensive evaluation index is calculated through a multi-dimensional comprehensive evaluation model, and dynamic correction and iterative optimization are carried out in combination with a feedback mechanism, ensuring the scientificity and effectiveness of the scheduling scheme, thereby improving the overall efficiency of emergency resource scheduling.

[0029] As a preferred embodiment of the technical solution of the present invention, the disaster environment data includes geospatial data and real-time disaster evolution prediction data; the emergency resource data includes the resource inventory, types, and locations of multi-level reserve warehouses, as well as the attributes and quantity of transport vehicles; the transportation network data includes the road network topology, traffic capacity, and dynamic damage probability.

[0030] The geospatial data refers to data describing the location, shape, and attributes of objects on the Earth's surface. This data can be acquired through satellite remote sensing imagery, drone aerial photography data, high-precision topographic maps, digital elevation models (DEMs), etc., or by integrating various map data, administrative division data, population density distribution data, etc., using a Geographic Information System (GIS) platform. The real-time disaster evolution prediction data refers to dynamic prediction information on the changes in the scope, intensity, and impact of a disaster over time after its occurrence. This data can be extrapolated based on natural disaster simulation and prediction systems such as meteorological models, hydrological models, and earthquake models, combined with real-time monitoring data, or by using big data analysis and machine learning algorithms to perform pattern recognition and trend prediction on historical disaster data and real-time sensor data.

[0031] The resource inventory, types, and locations of the multi-level reserve warehouses refer to the real-time inventory, specific types, and precise geographical coordinates of various materials (such as food, medicine, tents, medical equipment, etc.) distributed in emergency material warehouses at different administrative levels or geographical locations. Its function is to provide detailed information on deployable resources, ensuring that dispatch plans are based on actually available materials. This information can be obtained by establishing a unified emergency material management information system and updating the inbound and outbound records of reserve warehouses at all levels in real time. The attributes and quantities of transport vehicles refer to the carrying capacity, speed, fuel type, current location, availability, and total quantity of various vehicles (such as trucks, helicopters, and ships) that can be used to transport emergency materials. Its function is to assess transport capacity, select the most suitable transport methods and routes, and ensure that resources can be delivered efficiently. This information can be obtained by tracking the location and status of transport vehicles in real time through a vehicle management system (VMS) or GPS positioning system and recording their basic attributes.

[0032] The road network topology refers to the connectivity, geometric layout of nodes (intersections) and edges (road segments) of the road network, as well as static attributes such as road grade and length. Its function is to construct a basic traffic path model for path planning and traffic flow simulation. This structure can be obtained by extracting road centerline data and intersection data from high-precision electronic maps and GIS databases, and constructing a graph theory model. The capacity refers to the maximum number of vehicles that can pass through a road segment or intersection within a specific time period. Its function is to assess road carrying capacity, predict traffic congestion, and optimize route selection. This capacity can be estimated based on static parameters such as road type, number of lanes, speed limit, and traffic light timing, combined with a traffic engineering model. The dynamic damage probability refers to the likelihood that road segments or nodes in the road network will be damaged, interrupted, or have reduced capacity due to a disaster, and its trend over time. Its function is to introduce uncertainty factors, making the scheduling scheme more robust and able to cope with emergencies such as road interruptions. This probability can be assessed using probabilistic statistical models or expert systems, taking into account disaster type (such as earthquake intensity, flood depth), road structure type, and historical damage data.

[0033] Emergency resource data encompasses the resource reserves, types, and locations of multi-level reserve depots, as well as the attributes and quantities of transport vehicles. This ensures that dispatching plans can be meticulously planned based on actual available materials and transportation capacity, effectively avoiding blind and irrational resource allocation. Traffic network data incorporates road network topology, capacity, and dynamic damage probabilities, allowing scenarios to fully consider the static structure of the traffic network and dynamic risks under disaster impacts, such as road closures or capacity reductions. This enhances the adaptability and robustness of subsequent optimization models in complex and uncertain environments. This comprehensive data integration and scenario parameterization enable the entire emergency resource dispatching optimization simulation method to more closely resemble real disaster situations, generating more operational and reliable dispatching plans, significantly improving the scientific rigor and efficiency of emergency response.

[0034] As a preferred embodiment of the technical solution of the present invention, the emergency resource scheduling optimization model includes a multi-stage stochastic programming model and a robust optimization model. Its objective function includes at least minimizing the total response time, maximizing the fairness of demand satisfaction, and minimizing the total system cost. Its decision variables include resource allocation, transportation routes, and inventory dynamics.

[0035] In one example of the technical solution of this invention, the multi-stage stochastic programming model or robust optimization model is a mathematical tool for handling uncertain optimization problems. The multi-stage stochastic programming model considers multiple possible future scenarios and their probabilities to formulate the optimal decision for the current stage, aiming to achieve the desired performance under all scenarios. For example, the uncertainty evolution can be represented by constructing a scenario tree, and the Benders decomposition algorithm or L-shaped method can be used to solve the problem, decomposing the original problem into a deterministic main problem and stochastic subproblems, and iteratively solving to obtain the optimal solution. The robust optimization model aims to find a solution that maintains good performance even under the worst-case scenario. It captures the possible fluctuation range of parameters by defining an uncertainty set, thereby ensuring that the solution has strong resistance to parameter perturbations. For example, the uncertain parameters can be modeled as variables belonging to a certain convex set (such as an interval, ellipsoid, or polyhedron), and then the original problem can be transformed into a deterministic convex optimization problem, solved using the interior-point method or subgradient method, etc.

[0036] The objective function includes at least minimizing the total response time, maximizing the fairness of demand fulfillment, and minimizing the total system cost, aiming to achieve comprehensive optimization across multiple dimensions. Minimizing the total response time can be defined as the weighted sum of the times from the occurrence of the disaster to the arrival of the first batch of resources at all affected points, or the weighted sum of the times when resources are fully met at all affected points. The weights can be set according to the urgency or population density of the affected points. Maximizing the fairness of demand fulfillment can be measured by minimizing the variance or Gini coefficient of the resource fulfillment rate at each affected point, or by maximizing the lowest resource fulfillment rate among all affected points to ensure basic fairness. Minimizing the total system cost can include the sum of transportation costs, warehousing costs, labor costs, and other management and operating costs.

[0037] The decision variables include resource allocation, transportation routes, and inventory dynamics, which directly determine the specific execution details of the scheduling plan. Resource allocation can be represented as the type and quantity of resources allocated from which reserve to which disaster-stricken area at each time period. Transportation routes can be represented as a series of nodes and edges in the transportation network through which resources travel from the reserve to the disaster-stricken area, including the choice of road and transportation vehicle. Inventory dynamics can be represented as the changes in resource inventory levels at each reserve and disaster-stricken area at different time periods, including resource receipt, consumption, and replenishment.

[0038] By employing the aforementioned technical solutions and integrating the disaster environment data, emergency resource data, transportation network data, and historical disaster data gathered in step S1, the multi-stage stochastic programming model or robust optimization model proposed in this application can effectively capture the uncertainties brought about by the evolution of disasters and the dynamic probability of damage to transportation networks. By pre-considering multiple future scenarios or the worst-case scenario, it generates a more forward-looking and adaptive scheduling scheme. Simultaneously, by optimizing the three objectives of minimizing total response time, maximizing fairness in demand fulfillment, and minimizing total system cost, it avoids the uneven resource allocation or low economic efficiency that may result from optimizing a single objective, ensuring a comprehensive balance between timeliness, fairness, and economy in the scheduling scheme.

[0039] As a preferred embodiment of the technical solution of the present invention, the simulation environment based on multi-agent and geographic information system in step S3 includes: The environment layer is a static geospatial and transportation network model based on GIS. The entity layer includes intelligent agents at disaster sites, intelligent agents for transport vehicles, intelligent agents for resources, and intelligent agents for the command center. Each intelligent agent has autonomous behavior rules and interaction logic. The process layer simulates the dynamic processes of resource loading and unloading, transportation navigation, traffic flow, and information transmission delays.

[0040] In one embodiment of the technical solution of this invention, the environment layer is a static geospatial and traffic network model based on GIS. This environment layer aims to provide the physical foundation and spatial constraints of the simulation environment. GIS (Geographic Information System) is a system used to capture, store, manage, analyze, and display geospatial data, capable of accurately representing the geographical features of the real world. The static geospatial model can include relatively stable geographical features before a disaster occurs or during its evolution, such as terrain, landforms, water systems, and building distribution. The traffic network model details the topology, connectivity, number of lanes, speed limits, and other attributes of roads. In practical applications, existing geographic information databases (such as existing map service apps) can be used to obtain basic geographic data and road network data, which can then be preprocessed to construct vector or raster data models. By constructing this environment layer, the spatial constraints of the simulation environment are ensured to be consistent with those of the actual disaster scenario, providing an accurate foundation for subsequent resource scheduling, path planning, and traffic congestion simulation.

[0041] The entity layer comprises disaster-affected area agents, transport vehicle agents, resource agents, and command center agents, each with autonomous behavior rules and interaction logic. An agent is an autonomous software entity with perception, decision-making, and action capabilities, used to simulate various participants involved in disaster response and their dynamic behaviors and interactions. The disaster-affected area agent represents locations within the disaster area requiring rescue; its attributes may include resource demand, damage level, priority, etc., and it can issue requests based on changes in demand. The transport vehicle agent represents vehicles used to transport emergency supplies; its attributes may include load capacity, speed, current location, fuel consumption, etc., and it possesses route planning and dynamic adjustment capabilities. The resource agent represents the emergency supplies themselves; its attributes may include type, quantity, storage location, shelf life, etc. The command center agent is responsible for macro-level scheduling decisions, information dissemination, and coordination, such as adjusting scheduling strategies based on real-time conditions. The behavioral rules of these agents can be defined based on preset logical rules (Rule-based Agents), for example, the transport vehicle agent may follow the shortest path principle or avoid congested areas. By introducing these intelligent agents, the simulation system can simulate the complex dynamic behaviors of disaster-stricken areas, such as changing needs, vehicle route selection, resource allocation, and command and coordination.

[0042] The process layer simulates the dynamic processes of resource loading and unloading, transportation navigation, traffic flow, and information transmission delays. This layer aims to capture various dynamic and uncertain factors during resource scheduling, making the simulation results more realistic. The resource loading and unloading process involves the transfer of materials from storage depots to transport vehicles and from transport vehicles to disaster sites. Parameters such as average loading and unloading time and capacity can be set for simulation, and random disturbances can be introduced to reflect actual conditions. Transportation navigation is the process by which intelligent agents in transport vehicles select and adjust routes based on road network information and real-time traffic conditions. Path planning algorithms such as Dijkstra's algorithm can be used, combined with real-time traffic information for dynamic adjustments. Traffic flow simulates the movement of vehicles on roads, the formation and dissipation of congestion, and can be simulated using cellular automata models or microscopic traffic simulation models to simulate vehicle movement and interactions. Information transmission delays reflect the uncertainties of communication systems during disasters, such as information loss and extended transmission times. These can be simulated by setting parameters such as random delay times and packet loss rates.

[0043] In practical applications, the entity layer simulates the autonomous decision-making and complex interactions of intelligent agents at disaster sites, transport vehicles, resources, and command centers, reflecting the dynamic behaviors of changes in demand at disaster sites, route selection for transport vehicles, resource allocation, and command coordination in the real world. The process layer captures real-time changes and uncertainties in resource scheduling, such as resource loading and unloading efficiency, adjustments to transport navigation routes, traffic congestion, and information transmission delays, thereby significantly improving the dynamism and realism of the simulation.

[0044] As a preferred embodiment of the technical solution of the present invention, the process of calculating the comprehensive evaluation index in step S4 is as follows: The efficiency index, effectiveness index, fairness index, economic index, and economic index are weighted and summed to obtain the comprehensive evaluation index. The satisfaction threshold is connected to a dynamic adjustment port, which is used to adjust according to the disaster level and rescue stage.

[0045] The efficiency index is obtained as follows: The actual delivery time of resources to all disaster-stricken areas was statistically analyzed and compared with the expected delivery time for each area. Calculate the delay time at each point and normalize it based on the maximum allowable response time of the system. The overall efficiency index is obtained by averaging the timeliness scores of all disaster-affected locations. The effectiveness index is obtained as follows: The calculation requires summarizing the actual delivery quantities of all disaster-stricken areas and all types of resources, and comparing them with the actual demand of the corresponding areas. The smaller of the actual delivery quantity and the demand quantity is summed and then divided by the total demand quantity to obtain the effectiveness index. The fairness index is obtained as follows: The calculation is based on the distribution of resource satisfaction rate at each disaster site, and the Gini coefficient or Theil index is used to measure the degree of imbalance. The fairness index is determined based on the inverse ratio of the degree of imbalance. The robustness index is obtained as follows: In the simulation, a preset number of disturbance scenarios are simulated, the same scheduling scheme is executed respectively, and their overall performance is recorded; The robustness index is calculated by comparing the evaluation results under each disturbance scenario with the undisturbed baseline scenario. The economic indices are obtained as follows: The total cost of implementing the statistical scheme, including transportation expenses, warehousing expenses, labor expenses and management expenses, is compared with the upper and lower limits of costs in the simulation environment, and the economic index is obtained through linear normalization.

[0046] In one embodiment of the technical solution of the present invention, the multi-dimensional comprehensive evaluation model is used to calculate the comprehensive evaluation index, and its mathematical expression formula is as follows: ; in For efficiency weighting coefficients, This is the effect weighting coefficient. For fairness weighting coefficients, For robustness weighting coefficients, For economic weighting coefficients, ,and , , , as well as All greater than , It is an efficiency index that reflects the timeliness of resource delivery; This is an effectiveness index that reflects the adequacy of demand satisfaction. It is a fairness index that reflects the spatial fairness of resource allocation; It is a robustness index, reflecting the robustness of the solution in dealing with uncertainty; This is an economic index that reflects the economic cost efficiency of the scheduling process. This is a comprehensive evaluation index.

[0047] In one example of the technical solution of this invention, a comprehensive evaluation index is used. The efficiency index (E) is a single quantitative indicator that measures the overall performance of an emergency resource dispatch plan. It integrates multiple key performance dimensions through a weighted summation. The value of this index directly reflects the quality of the dispatch plan, providing an objective basis for subsequent plan selection and optimization. The efficiency index (E) aims to quantify the timeliness of resource delivery. It can be calculated based on the ratio of the average or median resource delivery time to a preset baseline time for all affected areas, or by statistically measuring the proportion of affected areas that complete resource delivery within the specified time. Effectiveness index... Used to assess the adequacy of demand fulfillment, it can be calculated as the ratio of the sum of the actual demand fulfilled at all disaster-stricken areas to the total demand, or it can be represented by the proportion of disaster-stricken areas that reach a preset demand fulfillment rate threshold to the total number of disaster-stricken areas. Fairness Index The focus is on the spatial balance of resource allocation, which can be assessed based on the differences in resource availability per unit population or unit affected area at various disaster-stricken points. This can be done by calculating the reciprocal of the variance or standard deviation, or by comparing the range or interquartile range of resource satisfaction rates at different disaster-stricken points. The robustness index R measures the robustness of a scheduling scheme in the face of uncertainty. It can be achieved by introducing random events (such as temporary road closures, fluctuations in resource depletion rates, and random increases in demand) into the simulation and statistically analyzing the average performance degradation of the scheme under these disturbances, or by calculating the ratio of the scheme's minimum performance in the worst-case scenario to its maximum performance in the ideal scenario. The economic index C is used to evaluate the economic cost efficiency of the scheduling process. It can be calculated by calculating the reciprocal of the ratio of total cost (including transportation costs, warehousing costs, and labor costs) to a preset budget ceiling, or by comparing actual costs with historical average costs or industry benchmark costs. Furthermore, the weights in the formula... , , , as well as The importance of each dimension index is determined by subjective and objective weighting methods such as expert scoring, analytic hierarchy process (AHP) or entropy weighting. These weights can also be dynamically adjusted based on factors such as the current disaster level, rescue stage, and resource scarcity. For example, efficiency and effectiveness may be emphasized in the initial rescue stage, while economy and equity may be emphasized in the recovery stage.

[0048] As a preferred embodiment of the present invention, the efficiency index E is obtained as follows: First, the actual resource delivery time of all disaster-stricken points is statistically analyzed and compared with the expected delivery time of each point; the delay time of each point is calculated (actual time minus expected time, if it is earlier, it is counted as 0), and then normalization is performed based on the maximum response time allowed by the system; finally, the average value of the timeliness scores of all disaster-stricken points is taken to obtain the overall efficiency index. Specifically, the efficiency index aims to quantify the timeliness of emergency resource delivery. Its acquisition first involves comparing the actual resource delivery time with the expected delivery time for all affected locations. Actual resource delivery time refers to the actual duration from resource dispatch to arrival at the affected location, which can be obtained through timestamps recorded in the simulation environment. Expected delivery time can be a preset ideal delivery time based on the urgency of the disaster, geographical location, resource type, etc. For example, different expected delivery times can be set according to the priority of the affected location or the resource type, or predictions can be made based on historical data and rules of thumb. Based on this, the delay time for each affected location is calculated, which is the actual delivery time minus the expected delivery time. If the actual delivery time is earlier than or equal to the expected delivery time, the delay time is counted as 0, indicating timely delivery. To convert the delay time into a comparable indicator, normalization is required using the system's maximum allowable response time as a benchmark. The maximum response time is the longest resource delivery time the system can tolerate under extreme conditions, which can be set according to national or industry emergency standards or determined through expert evaluation. Normalization can unify delay times of different dimensions into a range of 0 to 1, forming a timeliness score. Finally, by averaging the timeliness scores of all affected points, the overall efficiency index E is obtained. This averaging method can comprehensively reflect the overall timeliness level of resource delivery in the entire disaster area. Another approach is to use a weighted average method, assigning different weights based on the urgency or population density of the affected points to more accurately reflect the timeliness needs of key areas.

[0049] The effectiveness index F is obtained as follows: When calculating, the actual delivery quantity of all disaster-stricken points and all types of resources needs to be summarized and compared with the actual demand of the corresponding points; take the smaller value between the actual delivery quantity and the demand (i.e., to avoid oversupply being included), sum them up and divide by the total demand to obtain the effectiveness index. Specifically, the effectiveness index F measures the degree to which an emergency resource allocation plan meets the needs of disaster-stricken areas. Its calculation first requires summarizing the actual delivery quantities of all disaster-stricken areas and all resource types, and comparing them with the actual demand at each point. The actual delivery quantity refers to the specific quantity of various resources that ultimately arrive at the disaster-stricken area through simulation. The actual demand is the resource demand dynamically generated based on the disaster scenario and the behavioral logic of the agents at the disaster-stricken area. To avoid the interference of oversupply on the evaluation results, the smaller value between the actual delivery quantity and the demand quantity is taken. This means that even if the delivery quantity exceeds the demand quantity, only the demand-based portion is calculated, thus truly reflecting the effectiveness of demand fulfillment. For example, if a disaster-stricken area needs 100 units of supplies and actually delivers 120 units, it is counted as fulfilling 100 units; if it actually delivers 80 units, it is counted as fulfilling 80 units. The effectiveness index F is obtained by summing the smaller values ​​for fulfillment across all disaster-stricken areas and all resource types, and then dividing by the total demand across all disaster-stricken areas. The total demand is the sum of the demands for all resource types across all disaster-stricken areas.

[0050] The fairness index G is obtained by calculating the distribution of resource satisfaction rate (actual delivery / demand) at each disaster point, using the Gini coefficient or Theil index to measure the degree of imbalance, and then converting it into a fairness index by "1-degree of imbalance". Specifically, the equity index G aims to assess the balance of resource allocation across different disaster-stricken areas. It is calculated based on the distribution of resource satisfaction rates across these areas. The resource satisfaction rate is the ratio of actual delivery to demand at each disaster-stricken area, reflecting the degree to which a single area receives resources. To quantify the degree of imbalance in this distribution, this application uses the Gini coefficient or the Theil index. The Gini coefficient is a commonly used indicator for measuring income distribution fairness and is also applicable to measuring the fairness of resource allocation; its value ranges between 0 and 1, with values ​​closer to 0 indicating a fairer distribution. The Theil index is also an indicator for measuring inequality and is more sensitive to extreme values. These two indices can objectively reflect the differences in resource satisfaction rates across different disaster-stricken areas from a statistical perspective. Through a conversion of "1 - degree of imbalance," the degree of imbalance represented by the Gini coefficient or Theil index is transformed into the equity index G, such that a larger index value indicates a fairer distribution. For example, if the Gini coefficient is 0.3, the equity index is 0.7.

[0051] The robustness index R is obtained by simulating various disturbance scenarios (such as road interruption, sudden demand change, resource loss, etc.) in the simulation, executing the same scheduling scheme and recording its comprehensive performance; comparing the evaluation results under each disturbance scenario with the undisturbed baseline scenario, and calculating the robustness index. Specifically, ;in The number of perturbation scenarios. For the first A comprehensive evaluation index under various disturbances It is a comprehensive evaluation index under undisturbed conditions.

[0052] Among these scenarios, road disruptions simulate localized paralysis of the transportation network due to disasters; sudden demand fluctuations simulate a sudden increase or decrease in demand at disaster-stricken areas; and resource depletion simulates unexpected resource losses during transportation. Under each disturbance scenario, the same scheduling scheme is executed, and its overall performance is recorded, i.e., the comprehensive evaluation index under the current disturbance scenario. Simultaneously, a undisturbed baseline scenario is also required, i.e., the comprehensive evaluation index obtained by executing the scheduling scheme under ideal conditions (without any disturbances). The robustness index R is determined by evaluating the results under each disturbance scenario. Compared with the unperturbed baseline scenario The formula quantifies the ability of a scheme to maintain performance under different disturbances by calculating the average of the ratio of the comprehensive evaluation index under all disturbance scenarios to the baseline scenario index.

[0053] The economic index C is obtained by comparing the total cost of the statistical scheme execution, including various expenditures such as transportation, warehousing, manpower, and management, with the upper and lower limits of possible costs in the simulation environment, and obtaining the economic index through linear normalization. Specifically, ;in This represents the actual total cost (including transportation, inventory, labor, etc.). , These represent the highest and lowest possible costs in the simulation.

[0054] The economic index is obtained by statistically analyzing the total cost of implementing the plan, including transportation costs (such as fuel and vehicle wear and tear), warehousing costs (such as temporary warehouse rent and material storage), labor costs (such as rescue worker wages and overtime pay), and management costs (such as information system maintenance and dispatching expenses). These cost data can be calculated using parameters such as resource consumption and time consumption in the simulation environment. To convert the actual total cost into a comparable index, it needs to be compared with the upper and lower limits of costs that may occur in the simulation environment. Cost Upper Limit It can be set as the highest possible cost under the least economical scheduling scheme, and the lower limit of the cost. This can be set as the lowest possible cost under the most ideal and economical scheduling scheme. These upper and lower limits can be determined through historical data analysis, expert experience, or multiple simulation experiments. The actual total cost is then processed through linear normalization. Mapping to the range of 0 to 1, we obtain the economic index C, which is mathematically expressed as follows: This formula ensures that the lower the cost, the higher the economic index, thus intuitively reflecting cost efficiency. Another approach is to use logarithmic normalization or nonlinear normalization methods to adapt to different cost distributions, or to define the economic index as the cost-benefit ratio, i.e., the overall benefit that a unit cost can bring.

[0055] It should be noted that the above scheme actually provides a specific and quantitative calculation method for the multi-dimensional comprehensive evaluation model, thereby solving the problem that the evaluation results are highly subjective and cannot fully reflect the comprehensive effectiveness of the scheduling scheme. The quantitative calculation of the efficiency index E, by comparing the actual and expected delivery times and performing normalization, can objectively reflect the timeliness of resource delivery and avoid the influence of subjective judgment on the time efficiency assessment. The effectiveness index F, by accurately calculating the ratio of the actual met quantity to the demand quantity and avoiding the interference of oversupply, ensures an accurate assessment of the sufficiency of demand satisfaction, enabling the scheduling scheme to respond more accurately to the needs of the disaster area. The fairness index G uses mature statistical tools such as the Gini coefficient or the Theil index to transform the fairness of resource allocation from a qualitative description to a quantitative indicator, effectively avoiding human bias and promoting the balanced allocation of resources among different disaster-stricken points. The robustness index R, by simulating the scheme under various disturbance scenarios and comparing it with the baseline scenario, can comprehensively evaluate the robustness of the scheduling scheme in complex and uncertain environments, providing decision-makers with a quantitative basis for the scheme's risk resistance capability. The economic index C, by statistically analyzing the total cost and linearly normalizing it to the upper and lower limits of the cost, makes the assessment of cost efficiency more standardized and objective, helping to optimize resource utilization efficiency while ensuring the effectiveness of rescue efforts. The introduction of these specific quantitative indicators enables the comprehensive evaluation index I in step S4 to more accurately and comprehensively reflect the true performance of the scheduling scheme, thus providing reliable feedback for the dynamic correction in step S5 and the iterative optimization in step S6. This significantly improves the scientific rigor, adaptability, and overall rescue effectiveness of the simulation method for optimizing the scheduling of disaster emergency resources across the entire region.

[0056] Based on the above, the satisfaction threshold refers to a numerical limit used to measure whether the overall performance of an emergency resource dispatching plan has reached an acceptable level. When the comprehensive evaluation index calculated by the multi-dimensional comprehensive evaluation model is lower than this threshold, the system will trigger a feedback mechanism to correct the dispatching plan. The setting of this threshold directly affects the system's judgment on the merits of the plan and the frequency and timing of feedback corrections.

[0057] The disaster severity level refers to the classification of the severity, scope of impact, and urgency of a disaster event. It can be divided according to the disaster classification standards established by national or local emergency management departments, for example, classifying disasters into different levels such as extremely serious (Level I), serious (Level II), relatively serious (Level III), and general (Level IV). Alternatively, it can be based on real-time monitoring data, such as the number of casualties, the number of damaged houses, the distance of traffic disruptions, and the risk of secondary disasters, through expert systems or machine learning models for dynamic assessment and level determination.

[0058] The rescue phases refer to different periods or task priorities of emergency response actions on a timeline. For example, they can be divided according to the time progression after a disaster, such as 0-24 hours as the emergency response phase, 24-72 hours as the transitional rescue phase, and 72 hours as the recovery phase. Alternatively, they can be divided according to the nature and completion status of the rescue mission, such as the life search and rescue phase, the casualty treatment phase, the material transportation phase, and the infrastructure repair phase.

[0059] The dynamic adjustment refers to the fact that the satisfaction threshold is not fixed, but is adjusted in real time or periodically based on the current disaster level and rescue stage. One implementation method is to pre-establish a mapping table or rule base to map different combinations of disaster levels and rescue stages to corresponding satisfaction thresholds. For example, in the emergency response stage of a high-level disaster (such as Level I), to ensure a rapid response, the satisfaction threshold can be set relatively low, allowing the plan to be imperfect in the initial stage but to be quickly launched; while in the low-level disaster or recovery stage, the threshold can be set higher to pursue a more refined and optimized plan. Another implementation method is to use adaptive algorithms, such as fuzzy logic systems or reinforcement learning-based strategies, to learn and adjust the satisfaction threshold in real time based on historical data, expert experience, and the effects of current simulations, in order to better balance response speed, resource optimization, and demand satisfaction.

[0060] Through the above technical solutions, the satisfaction threshold is no longer a static, fixed value, but can be flexibly adjusted according to the actual situation of the disaster and the progress of rescue efforts. During periods of high disaster severity and urgent rescue missions, the satisfaction threshold can be appropriately lowered to trigger the feedback mechanism more quickly, ensuring timely response of emergency resources and avoiding delays in rescue efforts due to the pursuit of extreme optimization. Conversely, during periods of lower disaster severity or in the recovery phase, the satisfaction threshold can be increased, prompting the system to generate more refined and optimized scheduling plans to maximize resource utilization efficiency and demand satisfaction.

[0061] Figure 2A structural diagram of an emergency resource scheduling optimization simulation system for a comprehensive disaster situation is shown. In a preferred embodiment of the technical solution of the present invention, an emergency resource scheduling optimization simulation system for a comprehensive disaster situation is also provided, the system 10 comprising: The scenario parameterization module 11 is used to integrate disaster environment data, emergency resource data, transportation network data and historical disaster data to define and parameterize the disaster scenario across the entire domain. The scheduling scheme generation module 12 is used to establish an emergency resource scheduling optimization model containing spatiotemporal evolution process based on the scenario constructed by the scenario parameterization module, and to generate at least one preliminary scheduling scheme using the optimization algorithm. The scheme simulation module 13 is used to put the preliminary scheduling scheme generated by the scheduling scheme generation module into the simulation environment based on multi-agent and geographic information system to execute the simulation process, simulating resource flow, traffic congestion, dynamic changes in demand and unexpected events. Data analysis module 14 is used to construct a multi-dimensional comprehensive evaluation model, process the raw data output by the scheme deduction module simulation, calculate the comprehensive evaluation index, and trigger a feedback mechanism when the comprehensive evaluation index is lower than the preset satisfaction threshold. The model optimization module 15 is used to recursively optimize the emergency resource scheduling optimization model in the scheduling scheme generation module based on the evaluation results of the data analysis module; the optimization methods include adjusting the objective function weights and modifying the model parameters. The loop execution module 16 is used to loop through the scenario parameterization module to the model optimization module until the preset loop exit conditions are met, and output the final scheduling scheme and a complete simulation evaluation report; the loop exit conditions include the comprehensive evaluation index reaching the satisfaction threshold and the loop number reaching the preset maximum iteration number.

[0062] Furthermore, the disaster environment data includes geospatial data and real-time disaster evolution prediction data; the emergency resource data includes the resource inventory, types, and locations of multi-level reserve warehouses, as well as the attributes and quantity of transport vehicles; and the transportation network data includes the road network topology, traffic capacity, and dynamic damage probability.

[0063] Specifically, the emergency resource scheduling optimization model includes a multi-stage stochastic programming model and a robust optimization model. Its objective function includes at least minimizing the total response time, maximizing the fairness of demand satisfaction, and minimizing the total system cost. Its decision variables include resource allocation, transportation routes, and inventory dynamics.

[0064] Furthermore, the simulation environment based on multi-agent and geographic information systems in the scheme inference module includes: The environment layer is a static geospatial and transportation network model based on GIS. The entity layer includes intelligent agents at disaster sites, intelligent agents for transport vehicles, intelligent agents for resources, and intelligent agents for the command center. Each intelligent agent has autonomous behavior rules and interaction logic. The process layer simulates the dynamic processes of resource loading and unloading, transportation navigation, traffic flow, and information transmission delays.

[0065] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for simulation of optimization of emergency resource scheduling in a total disaster, characterized in that, The method comprises: S1: integrating disaster environment data, emergency resource data, traffic network data and historical disaster data, defining and parameterizing global disaster scenarios; S2: based on the scenario constructed in step S1, an emergency resource scheduling optimization model containing a time-space evolution process is established, and at least one preliminary scheduling scheme is generated using an optimization algorithm; S3: the preliminary scheduling scheme generated in step S2 is put into a simulation environment based on multi-agent and geographic information system to execute the deduction process, simulating resource flow, traffic congestion, demand dynamic change and unexpected events; S4: a multi-dimensional comprehensive evaluation model is constructed to process the original data output by step S3 simulation, and a comprehensive evaluation index is calculated, when the comprehensive evaluation index is lower than the preset satisfaction threshold, a feedback mechanism is triggered; S5: according to the evaluation result of step S4, the emergency resource scheduling optimization model in step S2 is recursively optimized; the optimization mode includes adjusting the weight of the objective function and modifying the model parameters; S6: steps S2 to S5 are executed in a loop until a preset loop exit condition is met, and a final scheduling scheme and a complete simulation evaluation report are output; the loop exit condition includes that the comprehensive evaluation index reaches the satisfaction threshold and the number of loops reaches the preset maximum iteration number.

2. The method of claim 1, wherein, The disaster environment data includes geographic space data and real-time disaster evolution prediction data; the emergency resource data includes the resource inventory, types, locations of multi-level reserves, and the attributes and quantities of transportation carriers; the traffic network data includes road network topology, traffic capacity and dynamic damage probability.

3. The method of claim 2, wherein, The emergency resource scheduling optimization model includes a multi-stage stochastic programming model and a robust optimization model, and the objective function at least includes minimizing the total response time, maximizing the demand satisfaction fairness, and minimizing the total system cost, and the decision variables include resource allocation, transportation path and inventory dynamics.

4. The method of claim 3, wherein, The simulation environment based on multi-agent and geographic information system in step S3 includes: An environment layer based on GIS static geographic space and traffic network model; An entity layer including disaster point agent, transportation vehicle agent, resource agent and command center agent, each agent having autonomous behavior rules and interaction logic; A process layer simulating resource loading and unloading, transportation navigation, traffic flow and information transmission delay dynamic process.

5. The method of claim 1, wherein, In step S4, the process of calculating the comprehensive evaluation index is: Weighted sum of efficiency index, effect index, fairness index, economic index and economic index to get comprehensive evaluation index; The satisfaction threshold circumscribes a dynamic adjustment port, and the dynamic adjustment port is used according to the disaster level and the rescue stage.

6. The method of claim 5, wherein, The efficiency index is obtained by: Statistical analysis of the actual resource delivery time of all disaster points, and comparison with the expected delivery time of each point; Calculate the delay time of each point, and normalize based on the maximum response time allowed by the system as the benchmark; Take the average of the timeliness scores of all disaster points to get the overall efficiency index; The effect index is obtained by: The actual delivery quantity of all disaster points and all resource categories is aggregated and compared with the actual demand quantity of the corresponding points during calculation; The smaller value between the actual delivery quantity and the demand quantity is obtained, and the sum is divided by the total demand quantity to obtain the effect index; The fairness index is obtained in the following manner: The fairness index is obtained based on the resource satisfaction rate distribution of each disaster point, and the Gini coefficient or the Theil index is used to measure the imbalance degree, and the fairness index is determined according to the inverse ratio of the imbalance degree; The robustness index is obtained in the following manner: In the simulation, a preset number of disturbance scenarios are simulated, the same scheduling scheme is executed respectively, and the comprehensive performance is recorded; The evaluation results in each disturbance scenario are compared with the benchmark scenario without disturbance, and the robustness index is calculated; The economic index is obtained in the following manner: The total cost in the scheme execution is calculated, including transportation cost, storage cost, labor cost and management cost, which is compared with the cost upper and lower limits in the simulation environment, and the economic index is obtained through linear normalization.

7. A simulation system for optimizing the allocation of emergency resources in the event of a disaster, characterized in that it comprises: The system comprises: A scenario parameterization module for integrating disaster environment data, emergency resource data, traffic network data and historical disaster data, defining and parameterizing the global disaster scenario; A scheduling scheme generation module for establishing an emergency resource scheduling optimization model containing a time-space evolution process based on the scenario constructed by the scenario parameterization module, and generating at least one preliminary scheduling scheme by using an optimization algorithm; A scheme deduction module for placing the preliminary scheduling scheme generated by the scheduling scheme generation module into a simulation environment based on multi-agent and geographic information system to execute a deduction process, simulating resource flow, traffic congestion, demand dynamic change and unexpected events; A data analysis module for constructing a multi-dimensional comprehensive evaluation model, processing the raw data output by the scheme deduction module, and calculating the comprehensive evaluation index, when the comprehensive evaluation index is lower than a preset satisfaction threshold, triggering a feedback mechanism; A model optimization module for recursively optimizing the emergency resource scheduling optimization model in the scheduling scheme generation module according to the evaluation results of the data analysis module; the optimization method includes adjusting the objective function weight and modifying the model parameters; A loop execution module for cyclically executing the scenario parameterization module to the model optimization module until a preset loop exit condition is met, outputting the final scheduling scheme and a complete simulation evaluation report; the loop exit condition includes that the comprehensive evaluation index reaches the satisfaction threshold and the number of iterations reaches a preset maximum iteration number.

8. The simulation system of claim 7, wherein, The disaster environment data includes geographic spatial data and real-time disaster evolution prediction data; the emergency resource data includes the resource inventory, category, location of multi-level storage, and the attributes and quantity of transportation carriers; the traffic network data includes road network topology, traffic capacity and dynamic damage probability.

9. The simulation system of claim 8, wherein, The emergency resource scheduling optimization model includes a multi-stage stochastic programming model and a robust optimization model, and the objective function at least includes minimizing the total response time, maximizing the demand satisfaction fairness, and minimizing the total system cost, and the decision variables include resource allocation, transportation path and inventory dynamics.

10. The simulation system of claim 9, wherein, The simulation environment based on multi-agent and geographic information system in the scheme deduction module comprises: An environmental layer, which is a GIS-based static geospatial and traffic network model; An entity layer, which includes disaster point agents, transport vehicle agents, resource agents, and command center agents, each having autonomous behavior rules and interaction logic; A process layer, which simulates dynamic processes of resource loading and unloading, transport navigation, traffic flow, and information transmission delay.

Images