Social distance-based hybrid disaster emergency evacuation path simulation planning method and system

Abstract

The application discloses a social distance-based mixed disaster emergency evacuation path simulation planning method and system, and belongs to the technical field of emergency management. The method is aimed at the concurrent scene of major infectious disease epidemic and natural disaster, and proposes a K shortest path capacity planning algorithm with social distance, including a K shortest path generation operator with node capacity constraint, a conflict road section identification operator and a capacity distribution operator with social distance. Under the condition of parameter determination, the evacuation scheme is quickly generated to realize the dynamic balance of evacuation efficiency and infection risk. Further, the above operators are integrated into a simulation optimization framework to construct an alternative scheme library through random sampling and budget allocation under the condition of parameter uncertainty, thereby improving the reliability and robustness of the scheme. The application can quickly develop a flexible and reliable evacuation strategy in a complex disaster environment, avoid local congestion and scheme failure, and enhance the multi-disaster collaborative response capability.

Description

Technical Field

[0001] This invention relates to the field of emergency management technology, and more specifically, to a simulation planning method and system for emergency evacuation routes in hybrid disasters based on social distancing. Background Technology

[0002] Emergency evacuation under mixed disasters presents a complex decision-making problem involving coordinating infection control and rapid evacuation when both epidemics and natural disasters occur simultaneously. The cumulative effects of such disasters exacerbate public safety risks; for example, maintaining social distancing during floods reduces evacuation efficiency, while overcrowding increases the risk of epidemic spread. Researching this issue is of significant practical importance for developing emergency strategies that balance dual risks and improving multi-disaster collaborative response capabilities, especially given the increasing frequency of extreme weather events and infectious diseases globally.

[0003] Hybrid disaster evacuation is often modeled as a mixed-integer programming (MIP) problem, with route selection, personnel allocation, and social distancing decisions as core variables. The objective function aims to minimize evacuation time and infection risk, and it involves complex factors such as multiple sources, multiple destinations, and multiple cycles. Solution methods mainly rely on precise algorithms such as branch and bound and cutting planes, as well as commercial solvers such as CPLEX, supplemented by dynamic programming to handle multi-cycle decisions. These methods can solve small to medium-scale problems, but they are not adaptable to nonlinear constraints and high-dimensional decision variables, and they rely on deterministic assumptions, making it difficult to cope with the uncertainty of disaster environments.

[0004] Existing methods face three major challenges: First, evacuation network parameters (such as path capacity and travel time) are difficult to predict due to disasters, and the uncertainty causes the model solution to deviate from reality, making the simulation planning method of emergency evacuation routes based on social distance, which relies on deterministic assumptions, ineffective. Second, the characteristics of integer variables and nonlinear constraints increase the complexity of the problem, and the computational cost of traditional algorithms increases exponentially with scale. Third, accurate algorithms are inefficient in large-scale scenarios and cannot meet the real-time requirements of emergency decision-making.

[0005] Therefore, there is an urgent need to design a simulation planning method and system for emergency evacuation routes in hybrid disasters based on social distancing in order to solve the above problems. Summary of the Invention

[0006] To address the complexity and uncertainty of emergency evacuation under mixed disasters, this invention proposes a simulation planning method and system for emergency evacuation routes in mixed disasters based on social distancing. Through an efficient and reliable heuristic-simulation optimization ensemble algorithm, collaborative optimization is performed to achieve efficient and reliable evacuation decisions. First, a simulation optimization framework is proposed, and an efficient heuristic algorithm—the K-shortest path capacity planning algorithm with social distancing—is developed. This algorithm is implemented through three operators: a K-shortest path generation operator with node capacity constraints, a conflict segment identification operator, and a conflict segment capacity allocation operator with social distancing. These operators are used to quickly generate satisfactory evacuation plans and determine social distancing levels under given parameter conditions. The social distancing decision and conflict segment capacity allocation are tightly coupled to form a collaborative optimization mechanism, effectively avoiding increased infection risk and evacuation delays caused by local path congestion, and dynamically balancing evacuation time and infection risk. Secondly, the developed K-shortest path capacity planning algorithm system with social distancing is integrated into the simulation optimization framework to form a simulation planning method for emergency evacuation routes in mixed disasters based on social distancing. Under uncertain parameter conditions, an alternative solution library is constructed to avoid solution failure caused by extreme path conditions and improve the effectiveness and reliability of path planning decisions under the concurrent conditions of compound disasters.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing includes the following steps:

[0009] S1. Design a K-shortest path generation operator with node capacity constraints. Under the constraints of network topology, node capacity, arc capacity and arc travel time, generate the first K feasible evacuation paths for each source node, and calculate the maximum flow and evacuation time of each path.

[0010] S2. Design a conflict segment identification operator. Based on the multi-source path set and the number of evacuees, identify the spatiotemporal overlap between different source node paths and generate a conflict segment set and node priority.

[0011] S3. Design a capacity allocation operator for conflict road segments with social distancing, couple different social distancing levels with capacity allocation of conflict road segments, allocate traffic to each source node path during the discrete planning period, and output evacuation schemes under different social distancing levels after completing capacity update, and finally select the social distancing level with the highest satisfaction score.

[0012] S4. Design a simulation optimization framework that integrates the above heuristic operators. Under the condition of parameter uncertainty, perform random sampling and budget allocation to generate and screen multiple alternative evacuation schemes, and finally obtain the optimal emergency evacuation route scheme.

[0013] As a preferred embodiment of the present invention, the K-shortest path generation operator with node capacity constraints includes the following sub-steps:

[0014] (1) Transform the multi-terminal problem into a single-terminal problem, introduce a super-terminal in the network, and connect all the terminals to the super-terminal through arcs with infinite capacity and zero travel time;

[0015] (2) Initialize the source node counter and path search counter, copy the original network, and set the node capacity and arc capacity;

[0016] (3) Use Dijkstra's algorithm to find the shortest path from the source node to the super endpoint in the network. When a path exists, prioritize the path with the shortest travel time and remaining flow.

[0017] (4) Calculate the maximum flow of the path, which is determined by the minimum node capacity and arc capacity on the path; calculate the total travel time of the path, which is the sum of the travel times of all arcs on the path;

[0018] (5) Update node capacity and arc capacity, and remove the node from the network when the capacity is zero;

[0019] (6) Repeat until the number of feasible paths to the source node reaches the preset threshold, or until there are no feasible paths.

[0020] As a preferred technical solution of the present invention, the conflict segment identification operator includes the following sub-steps:

[0021] (1) Calculate the potential evacuation time of each source node and sort them in descending order of value, giving higher priority to nodes with longer potential evacuation times;

[0022] (2) Traverse the path of each source node and check if there is any overlap with other paths at a certain arc segment;

[0023] (3) If a conflict exists, mark the arc segment and its path segment to the super endpoint as a conflict segment;

[0024] (4) Deduplicatize and aggregate all conflicting road segments to generate a global conflicting road segment set, and record its capacity and travel time information.

[0025] As a preferred embodiment of the present invention, the conflict segment capacity allocation operator with social distancing includes the following sub-steps:

[0026] (1) Input the discrete programming time period, the number of people to be evacuated from each source node, the path set, the maximum flow of the path, the path travel time, the set of conflict road segments, the social distancing level and its corresponding infection risk and capacity loss rate, and the weights;

[0027] (2) During the planning period, traffic is allocated according to the priority of nodes, and high-priority nodes are given priority to select their fastest path;

[0028] (3) When the path contains conflicting segments, the smaller value between the available flow at the current social distancing level and the remaining capacity of the conflicting segments shall be used as the allocated flow.

[0029] (4) When the path does not contain conflicting road segments, the available traffic under the current social distancing level is directly taken as the allocated traffic;

[0030] (5) After completing one flow allocation, update the conflict section capacity table until all source nodes are evacuated;

[0031] (6) Generate corresponding evacuation plans under multiple social distancing levels, and select the social distancing level with the highest score and its corresponding plan based on the weighted satisfaction score of evacuation time and infection risk.

[0032] As a preferred technical solution of the present invention, the social distancing level is a plurality of discrete levels, each level corresponding to a capacity reduction ratio and a relative infection risk value. The method generates evacuation plans at each level and compares and selects them based on satisfaction scores.

[0033] As a preferred technical solution of the present invention, the simulation optimization framework includes a candidate solution pre-selection step, specifically:

[0034] Parameter samples are randomly generated under a pre-selected budget. A K-shortest path generation operator with node capacity constraints is run to generate several evacuation schemes. High-frequency schemes are retained as a candidate scheme set by frequency statistics.

[0035] As a preferred technical solution of the present invention, the simulation optimization framework includes a sampling optimization step, specifically:

[0036] (1) Allocate an initial simulation budget for each candidate scheme, call the conflict segment identification operator and the conflict segment capacity allocation operator with social distance, and calculate the initial mean and variance of the scheme satisfaction score;

[0037] (2) During the sampling process, the posterior distribution information of the scheme is dynamically updated, and the optimal candidate scheme is selected for the next round of budget allocation based on the approximate value function results until all simulation budget allocations are completed.

[0038] As a preferred technical solution of the present invention, the simulation optimization framework includes a scheme selection step, specifically:

[0039] After sampling, the candidate solutions are ranked according to their posterior expected performance based on their satisfaction scores, and the top few best-performing solutions are selected as the final set of candidate solutions.

[0040] This invention proposes a multi-path planning system for emergency evacuation in hybrid disasters based on social distancing, comprising:

[0041] Parameter initialization module: used to input parameters such as network topology, node capacity, arc capacity, arc travel time, social distance level, and simulation budget;

[0042] Path generation module: Used to generate shortest path information for each source node under node capacity constraints;

[0043] Evacuation plan generation module: Used to generate evacuation plans by allocating capacity with social distancing based on route information and conflict segment identification results;

[0044] Alternative pre-generation module: used to generate a set of candidate evacuation schemes under uncertain parameter conditions;

[0045] Multi-option selection module: used to dynamically allocate and evaluate candidate options under simulation budget, and select the best performing alternative.

[0046] Social distancing level analysis module: used to simulate the effectiveness of solutions under different social distancing levels and output dynamic suggestions.

[0047] As a preferred technical solution of the present invention, the system can interface with the emergency command platform to realize real-time generation of evacuation routes, updating of plans, and recommendation of multiple plans under different mixed disaster scenarios.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] (1) Reliable decision support for multiple alternatives under uncertain environments. Existing simulation optimization methods mostly rely on fixed sampling strategies or empirical parameter settings, which are insufficiently adaptable to the dynamics of disaster evolution and parameter uncertainty. The social distancing-based simulation planning method for hybrid disaster emergency evacuation routes in this invention uses a Bayesian update mechanism to dynamically evaluate the expected performance of candidate alternatives under parameter distribution, and prioritizes the allocation of computing resources to potential alternatives, significantly enhancing the reliability of alternatives in the dynamic evolution of disasters. This method overcomes the limitations of low efficiency in traditional simulations and provides decision-makers with a multi-scenario alternative library that balances flexibility and reliability.

[0050] (2) Path generation capability with dynamic node capacity constraints. Traditional path planning algorithms usually only consider arc segment capacity constraints and ignore the impact of node capacity on evacuation efficiency, which can easily generate theoretically feasible but practically prone to local congestion. This invention uses a K-shortest path generation operator with node capacity constraints in a social distance-based hybrid disaster emergency evacuation path simulation planning method to dynamically adjust the remaining capacity of nodes. During the path generation stage, the capacity of nodes and arc segments is reduced simultaneously to ensure that the generated evacuation path not only meets the shortest time requirement, but also avoids the risk of systemic congestion caused by node overload. This significantly improves the feasibility of the path in actual evacuation scenarios. This step, as the basis of the heuristic algorithm, provides a prerequisite for reducing local congestion and improving the robustness of the solution.

[0051] (3) Efficient resolution mechanism for multi-source path conflicts. Existing methods mostly use global optimization models to handle path conflicts, which have high computational complexity and make it difficult to accurately locate bottleneck areas. The conflict segment identification operator of the social distance-based hybrid disaster emergency evacuation path simulation planning method of this invention quickly detects the spatiotemporal overlap of multi-source paths at key nodes and arcs, and clearly quantifies the resource competition relationship between paths. This step, as the key to the heuristic algorithm, can accurately identify high-load conflict segments, provide spatial dimension decision-making basis for subsequent dynamic capacity allocation, effectively reduce the computational burden of global optimization, further reduce the possibility of local congestion, and enhance the robustness of the scheme.

[0052] (4) Dynamic balance between infection risk and evacuation efficiency. Addressing the complex trade-off between path capacity loss and reduced infection risk caused by social distancing measures, the social distancing-based hybrid disaster emergency evacuation path simulation planning method of this invention employs a conflict segment capacity allocation operator with social distancing. This operator is tightly coupled with dynamic adjustment of social distancing, introduces heuristic rules, efficiently allocates evacuation capacity, and comprehensively compares different social distancing levels. This achieves an organic balance between minimizing evacuation time and controlling infection risk. Furthermore, this step further reduces local congestion and improves the robustness of the solution, making it a core element in achieving the technical effectiveness of the heuristic algorithm. Attached Figure Description

[0053] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute a limitation thereof.

[0054] Figure 1 This is a framework diagram of the multi-scheme emergency evacuation plan of the present invention;

[0055] Figure 2 This is a flowchart of the main simulation optimization ensemble algorithm of the present invention;

[0056] Figure 3This is a performance comparison chart of the method proposed in this invention with the OCBASS and EA methods. Detailed Implementation

[0057] The following will refer to the appendices in the embodiments of the present invention. Figure 1 The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the embodiments described below are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention. In order to provide a clearer explanation and description of the technical solutions and implementation methods of the present invention, specific examples of preferred implementations of the technical solutions of the present invention are introduced below.

[0058] A simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing includes:

[0059] Design a K-shortest path capacity planning algorithm with social distance, including three heuristic operators:

[0060] Operator 1 is a K-shortest path generation operator with node capacity constraints. Under the node capacity constraints, it generates the first K feasible evacuation paths for each source node and calculates their maximum flow and evacuation time. This operator can effectively avoid local congestion caused by node overload by simultaneously reducing the capacity of nodes and arc segments during the path generation stage.

[0061] Operator 2 is a conflict segment identification operator that identifies conflict segments between multi-source paths, providing support for subsequent personnel evacuation allocation. This operator rapidly detects the spatiotemporal overlap of multi-source paths at key nodes and arcs, accurately identifying high-load conflict segments and clearly quantifying the resource competition relationship between paths. This provides a precise spatial dimension decision-making basis for subsequent capacity allocation, reducing the possibility of localized congestion caused by resource competition.

[0062] Operator 3 is a capacity allocation operator for conflict sections with social distancing. It tightly couples the dynamic allocation of conflict section capacity with social distancing decisions, achieving synergistic optimization of evacuation time and infection risk. By efficiently allocating evacuation capacity and comprehensively comparing different social distancing levels, this operator achieves an organic balance between minimizing evacuation time and controlling infection risk, while avoiding local congestion and improving the robustness of the solution.

[0063] A simulation optimization framework is designed, and the above-mentioned heuristic operators are integrated into it to form a simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing. Under the condition of parameter uncertainty, the expected satisfaction score of the evacuation plan is estimated by simulation method, and then a highly reliable multi-alternative evacuation plan is generated.

[0064] Furthermore, the K-shortest path generation operator with node capacity constraints has the following sub-steps:

[0065] Under node capacity constraints, the first K feasible evacuation paths are generated for each source node, and their maximum flow and evacuation time are calculated. The specific process includes:

[0066] Input: network topology graph, node capacity, arc capacity, arc travel time, maximum number of paths.

[0067] Output: The set of paths for each source node, the maximum flow of each path, and the path travel time.

[0068] The problem is transformed into an equivalent single-terminal problem: Add a super terminator to the network with infinite capacity; select all other terms and add a directed arc to each terminator, setting their travel time to 0 and their arc capacity to infinite; add the newly added super terminator and its associated edges to the original network to obtain a new network; thus, the original problem is transformed into an equivalent single-terminal problem.

[0069] Algorithm initialization includes initializing the source node counter, initializing the path search counter, replicating the network, setting the initial node capacity, and setting the arc capacity.

[0070] Path search:

[0071] For each source node, Dijkstra's algorithm is used to find the shortest path to the super endpoint in the network, prioritizing the path with the shortest travel time and remaining flow.

[0072] If the path does not exist, terminate the search for the source node, record the actual number of paths generated, update the source node counter and the path search counter, and proceed to the path search step for the next source node.

[0073] Path flow and travel time calculation:

[0074] The maximum flow of a path is determined by the minimum node capacity and arc capacity on the path; the total travel time of a path is the sum of the travel times of all arcs on the path.

[0075] Network update:

[0076] Each time a new path is obtained, the capacity of the nodes on the path is updated: for each node passed through, its current capacity is reduced by the path flow. If the capacity of the node is 0 at this time, it is removed from the network along with all associated arcs.

[0077] Update the capacity of arc segments on the path: For each arc segment traversed, subtract the path traffic from its current capacity. If the capacity of the arc segment is 0 at this time, remove it from the network.

[0078] When moving to the next source node iteration, the network is reset to the original network.

[0079] Iteration Termination: If the number of paths obtained by the source node has not reached the threshold, return to the path search step; otherwise, reset the network and proceed to the next source node iteration process; the algorithm terminates when all source nodes have completed the path search iteration.

[0080] This step, as an important component of the heuristic algorithm, dynamically adjusts the remaining capacity of nodes and synchronously reduces the capacity of nodes and arcs during the path generation stage. This effectively avoids local congestion caused by node overload, lays the foundation for subsequent steps, and improves the robustness of the entire solution.

[0081] Furthermore, the conflict segment identification operator has the following sub-steps:

[0082] Identifying conflicting road segments among multiple routes supports subsequent evacuation personnel allocation. The specific process includes:

[0083] Input: Number of people to be evacuated from all source nodes, set of paths, path flow, and path travel time.

[0084] Output: The set of conflicting road segments for each path, and the set of non-repeating conflicting road segments.

[0085] Priority sorting:

[0086] Calculate the expected evacuation time for each source node, sort the source nodes in descending order, and assign higher priority to nodes with longer expected evacuation times.

[0087] Conflict detection:

[0088] For each path of each source node, traverse each path of other source nodes and compare them to find conflicting arcs;

[0089] If two paths coincide at a certain arc segment, then the segment consisting of that arc segment and all subsequent arc segments up to the super endpoint is marked as a conflict segment of the source node path.

[0090] Conflict section aggregation:

[0091] Remove duplicates from all conflicting road segments, generate a global set of conflicting road segments, and record the capacity and travel time of each conflicting road segment.

[0092] This step, as a key component of the heuristic algorithm, rapidly detects the spatiotemporal overlap of multi-source paths at key nodes and arcs, accurately identifies high-load conflict segments, and clearly quantifies the resource competition relationship between paths. This provides a precise spatial dimension decision-making basis for subsequent capacity allocation, reduces the possibility of local congestion caused by resource competition, and further enhances the robustness of the solution.

[0093] Furthermore, the conflict segment capacity allocation operator with social distancing has the following sub-steps:

[0094] Dynamically allocating conflict zone capacity to determine social distancing levels and balancing evacuation time with infection risk is a step closely coupled with dynamic social distancing adjustments. The specific process includes:

[0095] Input: Discrete planning time period, number of people to be evacuated at each source node, set of paths, maximum flow rate of each path, path travel time, set of conflict sections, social distancing level and its corresponding infection risk and capacity loss rate, and weights.

[0096] Output: Optimal social distancing plan, optimal evacuation plan.

[0097] Priority scheduling: Processing is done in descending order of source node evacuation time, prioritizing the allocation of faster path traffic to higher priority nodes.

[0098] Conflicting segment capacity allocation: When a path with conflicting segments can reach its destination within the current time period, the flow allocation value for that path during that time period is the smaller of its available flow at the current social distancing level and the current capacity of the conflicting segments it traverses. Because available flow changes with different social distancing levels, it affects the capacity allocation of conflicting segments. Conversely, the capacity status of conflicting segments prompts a reassessment of the social distancing level, creating a coupling relationship.

[0099] Conflict-free path capacity allocation: When a conflict-free path can reach its destination in the current time period, the flow allocation value for that path in that time period is the available flow under the current social distancing level.

[0100] After capacity allocation, the current source node needs to be determined to complete the evacuation: when the allocated traffic is greater than the number of people to be evacuated from the current node, the smaller of the two values ​​is taken as the traffic allocation value for that path in that time period, and the iteration process of the next source node is started after the capacity update of the conflict section is completed.

[0101] Conflict segment capacity update: After each conflict segment traffic is completed, update the remaining capacity of the conflict segment for each time period.

[0102] Social distancing selection: For each social distancing level, obtain the evacuation plan through the above steps, calculate the multi-objective weighted satisfaction score of evacuation efficiency and risk control, select the highest score as the optimal social distancing, and output the evacuation plan under the optimal social distancing level.

[0103] This step, as the core of the heuristic algorithm, is coupled with the dynamic adjustment of social distancing to efficiently allocate evacuation capacity and comprehensively compare different social distancing levels. This achieves an organic balance between minimizing evacuation time and controlling infection risk, while avoiding local congestion and improving the robustness of the solution.

[0104] Furthermore, the simulation optimization framework integrating heuristic operators has the following sub-steps:

[0105] Under conditions of parameter uncertainty, the expected satisfaction score of evacuation plans is estimated through simulation methods, thereby generating multiple highly reliable alternative evacuation plans. Specific steps include:

[0106] Inputs: network parameter distribution information, pre-selected budget, simulation budget, initial budget, number of pre-selected schemes, number of best alternative schemes.

[0107] Output: The set of optimal alternative solutions and the estimated expected satisfaction level of the solutions.

[0108] Candidate scheme pre-selection: Under the pre-selected budget, parameter samples are randomly generated, and the K shortest path generation operator with node capacity constraints is run to generate several path schemes. High-frequency schemes are retained as the candidate scheme set through frequency statistics.

[0109] Sampling optimization:

[0110] Initialization: Allocate an initial simulation budget for each candidate solution, and calculate the initial mean and variance of the satisfaction score using the conflict segment identification operator and the conflict segment capacity allocation operator with social distancing.

[0111] Dynamic sampling: After each round of sampling, the posterior distribution information is updated, and the next sampling scheme is dynamically selected based on the approximate value function until all simulation budgets are allocated.

[0112] Solution selection: After the simulation is completed, the solutions are sorted in descending order of expected satisfaction and posterior performance, and the best performing solutions are selected as the final output and included in the alternative solution library.

[0113] Furthermore, the present invention provides a multi-path planning system for emergency evacuation in hybrid disasters based on social distancing, comprising:

[0114] The parameter initialization module is used to initialize algorithm parameters, including network topology, node capacity, edge capacity and travel time, social distance level, simulation budget, etc.

[0115] The path generation module is used to generate the shortest path information for each source node based on the given or simulated network information scenario.

[0116] The evacuation plan generation module is used to identify conflicting road sections and dynamically allocate traffic based on route information, generating the optimal social distancing level and evacuation planning scheme.

[0117] The alternative pre-generation module is used to dynamically generate several candidate evacuation route planning schemes based on known uncertain parameter information.

[0118] The multi-option selection module is used to allocate budgets, evaluate and rank pre-generated alternative options, and finally select the best-performing alternative option to be included in the option library.

[0119] The social distancing level analysis module is used to simulate the choice of social distancing level under different scenarios and its impact on the effectiveness of evacuation plans, and to provide dynamic suggestions for the optimal social distancing level.

[0120] Example 1

[0121] like Figure 1 As shown, this embodiment discloses a K-shortest path capacity planning algorithm with social distancing, which is an important component of the social distancing-based hybrid disaster emergency evacuation route simulation planning method proposed in this invention, and includes the following steps:

[0122] S1: Provides a K-shortest path generation operator with node capacity constraints. Given network path information and node capacity constraints, it generates a feasible evacuation path for each source node and calculates its maximum flow and evacuation time.

[0123] S2: Provides a conflict segment identification operator to identify potential conflict segments between different source nodes and different paths;

[0124] S3: Provides a capacity allocation operator for conflict road segments with social distancing. Based on path and conflict road segment information, combined with dynamic adjustment of social distancing, it selects the optimal social distancing level and generates an evacuation plan.

[0125] Next, a detailed description will be given of a K-shortest path capacity planning algorithm with social distance disclosed in this embodiment.

[0126] In S1, based on the K-shortest path generation operator with node capacity constraints, for each source node, its... Find the shortest path and calculate its maximum flow and evacuation time. Given the network input... and node capacity Arc capacity Travel Time with Arc .

[0127] S101: Transform a multi-destination network problem into a single-destination network problem:

[0128] S1011: Introducing Virtual Destination Nodes ;

[0129] S1012: Move the original destination set Each node in the middle is pointed to by a new one-way arc. And set the arc travel time to 0 and the capacity to infinity;

[0130] S1013: Generate a new network This is the new single-destination network problem.

[0131] S102: For each source node, obtain its... Find the shortest path, calculate its maximum flow and evacuation time, and update the network. The following is the specific process for obtaining each shortest path for each source node:

[0132] S1021: For the source node Towards the destination The Shortest path On the Internet In this process, Dijkstra's algorithm is used to obtain the shortest path. If no shortest path exists, then let the source node... The actual available shortest path is Then let And move on to the next node. ;

[0133] S1022: Calculation Path Maximum flow : the minimum of the flow of all its arcs and the capacity of all the nodes it passes through, and let The remaining number of nodes passed and arcs minus ;

[0134] S1023: Calculation Path travel time :Will The travel time of all arcs is summed up to obtain the result;

[0135] S1024: Update nodes in the network: For nodes with 0 remaining capacity, remove them and the arcs directly connected to them from the network;

[0136] S1025: Update arcs in the network: For arcs with a remaining capacity of 0, remove them from the network;

[0137] S1026: Path Loop and Termination: If at this time... Then let the source node The actual available shortest path is , and then Proceed to the next node and make the network If at this time Then let Set the updated network as the new one. Then, proceed to the next path of that node in a loop;

[0138] S1027: Node Cyclicity and Termination: When ,and The algorithm terminates when there is no shortest path available.

[0139] S103: Algorithm Output: Output for each source node Shortest path and its maximum flow With evacuation time .

[0140] In S2, a conflict segment identification operator is used to identify potential conflict segments between different paths from different source nodes. The input is the number of people to be evacuated from each source node. , Shortest path and its maximum flow With evacuation time .

[0141] S201: Obtain source node priority:

[0142] S2011: Calculate the potential evacuation time for each node using the following formula. :

[0143] ;

[0144] S012: Arrange the potential evacuation times from largest to smallest to obtain a new order. Larger source nodes have higher priority.

[0145] S202: Starting with the fastest path from the source node with the highest priority, obtain the conflict path table for each path of each source node. :

[0146] S2021: With any node (include any path of itself Compare the paths to check for conflicts.

[0147] S2022: If a conflict exists, record it as Write .

[0148] S203: Summary of all Non-repeating conflicting road segments are included in a new set. .

[0149] S204: Output: Set of non-repeating conflicting road segments Node priority order Potential evacuation time Table of all conflicting paths .

[0150] In S3, a conflict segment capacity allocation operator with social distancing is used, combined with dynamic adjustment of social distancing, to allocate evacuees to various paths within the planning period and determine the social distancing level. The input is a set of non-repeating conflict segments. Node priority order Potential evacuation time Table of all conflicting paths The number of people to be evacuated from each source node , Shortest path and its maximum flow With evacuation time social distancing capacity reduction factor and infection risk coefficient Upper and lower limits of evacuation time Relative infection risk upper and lower limits Weight .

[0151] S301: Initialize the conflict path capacity table For each social distancing level, initialize its evacuation plan. and conflict zone capacity table This excludes the capacity reduction caused by social distancing.

[0152] S302: Obtain evacuation planning options at each social distancing level.

[0153] S303: Sort all source nodes in order of priority. Allocation is done in turn, completing the calculation of one node before moving on to the next:

[0154] S3031: For a given source node During a certain period Next, determine the path in sequence. If the destination cannot be reached at this moment, then determine the next path; if the destination can be reached, then allocate capacity.

[0155] S3032: Capacity Allocation: First, directly allocate the maximum throughput after deducting capacity reduction, then write... Then, it determines whether it has passed through any conflict section. If there is no conflict, it performs an evacuation completion judgment; if there is a conflict, it performs a conflict capacity calculation.

[0156] S3033: Conflict Capacity Calculation: Path Allocated capacity and The capacity of the conflicting road sections at this moment is compared, the minimum value is selected, and then the evacuation is completed.

[0157] S3034: Evacuation Completion Judgment: If The number of remaining unrelocated residents has not been reached; only the capacity table for conflict zones needs to be updated. Continue calculating the next path; if If the number exceeds the current number of people who have not yet been evacuated, then the minimum of the two values ​​will be taken as the evacuation allocation number. Then update the conflict segment capacity table. And terminate the calculation of the current source node and move on to the next source node;

[0158] S3035: Update the conflict segment capacity table: capacity of all relevant conflict paths at this moment. deduct ;

[0159] S3036: Obtain the social distancing level when all source node calculations are complete. The following evacuation plan And determine the evacuation completion time. .

[0160] S304: After completing evacuation planning at all social distancing levels, calculate the optimal satisfaction score using the following formula. And select the optimal level of social distancing :

[0161]

[0162] S305: Output: Best Satisfaction Score Optimal social distancing level Optimal evacuation plan Optimal evacuation time .

[0163] To demonstrate the effectiveness of the solution described in this embodiment, relevant experiments were conducted in conjunction with specific embodiments:

[0164] Set up a 62-node evacuation network, including 12 source nodes and 3 destination nodes. First, add a superdestination to transform the problem into a single-destination problem. Set three social distancing levels: unlimited, at least one meter, and at least two meters, denoted as [insert values ​​here]. The relative infection risks are 1.0, 0.4, and 0.3, respectively, and the capacity loss ratios are 0, 0.2, and 0.5, respectively. Assume the capacity of the arc, travel time, and node capacity are known. To solve the problem, only three steps, S1, S2, and S3, are needed to obtain an accurate and satisfactory solution. Experimental results show that the optimal social distancing level is one meter (…). The total evacuation time was 22 minutes, and the final satisfaction score was... After repeated experiments, the average running time of the algorithm was found to be 0.0660 seconds, demonstrating its computational efficiency.

[0165] Example 2

[0166] S4: Provides a simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing. By integrating three heuristic operators of the K-shortest path capacity planning algorithm with social distancing, it can quickly and efficiently generate multiple alternative solutions and incorporate them into the solution library under uncertain conditions.

[0167] In S4, such as Figure 2 As shown, within a simulation optimization framework, by integrating the above heuristic operators, multiple alternative solutions are generated under uncertain conditions, and the solutions are evaluated based on their satisfaction scores. Sort them by their expected values ​​to obtain the top The best alternative is included in the solution library. Input: Parameter information set Pre-selected budget for the plan Number of pre-selected schemes Total simulation budget Initial simulation budget of the scheme Final number of alternative solutions .

[0168] S401: In cases containing information on the distribution of uncertain parameters Next, perform uncertain parameters. In each random sampling, a set of evacuation route planning schemes is obtained and counted using step S1. If a scheme appears repeatedly, its count is incremented by 1. After the iteration is complete, the schemes are sorted in descending order of their frequency of occurrence, and the top-ranked schemes are retained. The options are selected as preliminary results. A satisfaction score is then obtained for each option. Prior expectation and variance Assume the expected value is unknown, but the variance is known. .

[0169] S402: Scheme Initialization: Pre-allocate resources for each scheme. A simulation budget is used to obtain initial observations;

[0170] S403: Simulation Budget Allocation

[0171] S4031: Calculate the posterior performance of each option using the following formula:

[0172]

[0173] in, The solution in this iteration The posterior mean and variance of satisfaction scores. For the plan The number of simulation budgets already allocated, This represents the average of the observed satisfaction scores under the simulation budget.

[0174] S4032: Calculate the approximate value function using the following formula:

[0175]

[0176] S4033: Select the alternative with the highest approximate value function value. Assign a simulation budget to it: incorporating information on the distribution of uncertain parameters Next, perform one random sampling of the uncertain parameters, and then for the alternative schemes... The planning scheme uses steps S2 and S3 to obtain a satisfaction score. Then, the posterior expected value is recalculated, thus completing one simulation.

[0177] S4034: Termination judgment: If at this time If there is still a remaining simulation budget, then return to step S4031 to continue the algorithm; if at this time... If the simulation budget is fully allocated, the loop exits and the result is output.

[0178] S404: Sort the alternatives in descending order based on the posterior expected value of their satisfaction scores, starting with the top three options. The selected option is the optimal alternative. Output: Expected satisfaction score for all alternative options. Best alternative solutions .

[0179] To demonstrate the effectiveness of the solution described in this embodiment, relevant experiments were conducted in conjunction with specific embodiments:

[0180] Based on the evacuation network in the previous specific embodiment, the capacity and travel time of the arcs and the capacity of the nodes are uncertain, but the mean and variance of their normal distribution are known, denoted as the parameter information set. Under conditions of uncertainty, step S4 is used to seek multiple optimal alternatives. The budget for each alternative is pre-selected. Number of pre-selected schemes Total simulation budget Initial simulation budget of the scheme Final number of alternative solutions Experimental results show that among the five pre-selected alternatives... middle, , , This is the best alternative.

[0181] To verify the accuracy of the solution, 10,000 simulations were performed independently for each candidate scheme, and the average score was used as the expected performance. The results were 0.5517, 0.5490, 0.5518, 0.5517, and 0.5491, indicating that schemes 1, 3, and 4 have higher actual expected performance, thus proving the accuracy of the algorithm. Furthermore, after repeated experiments, the average running time of the algorithm was found to be 188.8696 seconds, and accurate results were consistently obtained, demonstrating its computational efficiency and reliability.

[0182] To further verify the superiority of the proposed method, the simulation planning method for hybrid disaster emergency evacuation routes based on social distancing proposed in this invention is compared with the EA and OCBASS methods. The experimental setup is as follows. Based on the five pre-selected alternative schemes obtained in this embodiment, the simulation budget... Starting from [date], gradually increase the budget to [amount]. .make In each iteration, three methods—social distancing-based hybrid disaster emergency evacuation route simulation planning, EA, and OCBASS—are each independently executed 10,000 simulations. The number of times the optimal alternative is correctly selected is counted, and the probability of correct selection (PCS) for each method is obtained. Figure 3 As shown, the simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing has a significantly higher probability of correct selection than EA and OCBASS under the same simulation budget, demonstrating the advanced nature and efficiency of this invention.

[0183] Example 3:

[0184] This invention provides a multi-path planning system for emergency evacuation in hybrid disasters based on social distancing, comprising:

[0185] The parameter initialization module is used to initialize algorithm parameters, including network topology, node capacity, edge capacity and travel time, social distance level, simulation budget, etc.

[0186] The path generation module is used to generate the shortest path information for each source node based on the given or simulated network information scenario.

[0187] The evacuation plan generation module is used to identify conflicting road sections and dynamically allocate traffic based on route information, generating the optimal social distancing level and evacuation planning scheme.

[0188] The alternative pre-generation module is used to dynamically generate several candidate evacuation route planning schemes based on known uncertain parameter information.

[0189] The multi-option selection module is used to allocate budgets, evaluate and rank pre-generated alternative options, and finally select the best-performing alternative option to be included in the option library.

[0190] The social distancing level analysis module is used to simulate the choice of social distancing level under different scenarios and its impact on the effectiveness of evacuation plans, and to provide dynamic suggestions for the optimal social distancing level.

[0191] The modules involved in the above embodiment two correspond to those in embodiment one. For specific implementation details, please refer to the relevant description section of embodiment one.

[0192] The contents not described in detail in this description are existing technologies known to those skilled in the art. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing, characterized in that, Includes the following steps: S1. Design a K-shortest path generation operator with node capacity constraints. Under the constraints of network topology, node capacity, arc capacity and arc travel time, generate the first K feasible evacuation paths for each source node, and calculate the maximum flow and evacuation time of each path. S2. Design a conflict road segment identification operator. Based on the multi-source path set and the number of evacuees, identify the spatiotemporal overlap between paths from different source nodes, and generate a conflict road segment set and node priorities. The conflict road segment identification operator includes the following sub-steps: (1) Calculate the potential evacuation time of each source node and sort them in descending order of value, giving higher priority to nodes with longer potential evacuation times; (2) Traverse the path of each source node and check if there is any overlap with other paths at a certain arc segment; (3) If a conflict exists, mark the arc segment and its path segment to the super endpoint as a conflict segment; (4) Remove duplicates from all conflicting road segments and aggregate them to generate a global set of conflicting road segments, and record their capacity and travel time information; S3. Design a capacity allocation operator for conflict road segments with social distancing, coupling different social distancing levels with capacity allocation of conflict road segments. Distribute traffic to each source node path within a discrete planning time period, and output evacuation schemes under different social distancing levels after capacity updates are completed. Finally, select the social distancing level with the highest satisfaction score. The capacity allocation operator for conflict road segments with social distancing includes the following sub-steps: (1) Input the discrete programming time period, the number of people to be evacuated from each source node, the path set, the maximum flow of the path, the path travel time, the set of conflict road segments, the social distancing level and its corresponding infection risk and capacity loss rate, and the weights; (2) During the planning period, traffic is allocated according to the priority of nodes, and high-priority nodes are given priority to select their fastest path; (3) When the path contains conflicting road segments, the lower limit of the available flow under the current social distancing level and the remaining capacity of the conflicting road segments shall be used as the allocated flow. (4) When the path does not contain conflicting road segments, the available traffic under the current social distancing level is directly taken as the allocated traffic; (5) After completing one flow allocation, update the conflict section capacity table until all source nodes are evacuated; (6) Generate corresponding evacuation plans under multiple social distancing levels, and select the social distancing level with the highest score and its corresponding plan based on the weighted satisfaction score of evacuation time and infection risk. S4. Design a simulation optimization framework that integrates the K-shortest path generation operator with node capacity constraints, the conflict segment identification operator, and the conflict segment capacity allocation operator with social distancing. Under the condition of parameter uncertainty, random sampling and budget allocation are performed to generate and screen multiple alternative evacuation schemes, and finally obtain the optimal emergency evacuation route scheme.

2. The simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing as described in claim 1, characterized in that, The K-shortest path generation operator with node capacity constraints includes the following sub-steps: (1) Transform the multi-terminal problem into a single-terminal problem, introduce a super-terminal in the network, and connect all the terminals to the super-terminal through arcs with infinite capacity and zero travel time; (2) Initialize the source node counter and path search counter, copy the original network, and set the node capacity and arc capacity; (3) Use Dijkstra's algorithm to find the shortest path from the source node to the super endpoint in the network. When a path exists, prioritize the path with the shortest travel time and remaining flow. (4) Calculate the maximum flow of the path, which is determined by the minimum node capacity and arc capacity on the path; calculate the total travel time of the path, which is the sum of the travel times of all arcs on the path; (5) Update node capacity and arc capacity, and remove the node from the network when the capacity is zero; (6) Repeat until the number of feasible paths to the source node reaches the preset threshold, or until there are no feasible paths.

3. The simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing as described in claim 1, characterized in that, The social distancing level is divided into multiple discrete levels, each level corresponding to a capacity reduction ratio and a relative infection risk value. The method generates evacuation plans at each level and compares and selects them based on satisfaction scores.

4. The simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing as described in claim 1, characterized in that, The simulation optimization framework includes a candidate solution pre-selection step, specifically: Parameter samples are randomly generated under a pre-selected budget. A K-shortest path generation operator with node capacity constraints is run to generate several evacuation schemes. High-frequency schemes are retained as a candidate scheme set by frequency statistics.

5. The simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing as described in claim 1, characterized in that, The simulation optimization framework includes a sampling optimization step, specifically: (1) Allocate an initial simulation budget for each candidate scheme, call the conflict segment identification operator and the conflict segment capacity allocation operator with social distance, and calculate the initial mean and variance of the scheme satisfaction score; (2) During the sampling process, the posterior distribution information of the scheme is dynamically updated, and the optimal candidate scheme is selected for the next round of budget allocation based on the approximate value function results until all simulation budget allocations are completed.

6. The simulation planning method for emergency evacuation routes in hybrid disasters based on social distancing as described in claim 1, characterized in that, The simulation optimization framework includes a scheme selection step, specifically: After sampling, the candidate solutions are ranked according to their posterior expected performance based on their satisfaction scores, and the top few best-performing solutions are selected as the final set of candidate solutions.

7. A multi-path planning system for emergency evacuation in hybrid disasters based on social distancing, characterized in that, The method applied to any one of claims 1-6, comprising: Parameter initialization module: used to input network topology diagram, node capacity, arc capacity, arc travel time, social distance level, and simulation budget parameters; Path generation module: Used to generate shortest path information for each source node under node capacity constraints; Evacuation plan generation module: Used to generate evacuation plans by allocating capacity with social distancing based on route information and conflict segment identification results; Alternative pre-generation module: used to generate a set of candidate evacuation schemes under uncertain parameter conditions; Multi-option selection module: used to dynamically allocate and evaluate candidate options under simulation budget, and select the best performing alternative. Social distancing level analysis module: used to simulate the effectiveness of solutions under different social distancing levels and output dynamic suggestions.

8. The system as described in claim 7, characterized in that, The system can interface with the emergency command platform to enable real-time generation of evacuation routes, updating of plans, and recommendation of multiple plans under different mixed disaster scenarios.

Images