Multi-algorithm combinatorial optimization 120 emergency center ambulance task allocation method

Abstract

The invention discloses a 120 emergency center ambulance task allocation method based on multi-algorithm combinatorial optimization. The method comprises the following steps: S1, acquiring specific positions and illness conditions of all patients to be treated, specific positions of all idle ambulances, specific positions of all hospitals and the treatment capability of each patient in a current time period; s2, constructing a task allocation evaluation function of the ambulance and a treatment effect evaluation function obtained by each patient; s3, path planning is carried out in combination with an A * algorithm, and it is ensured that the ambulance selects the optimal path; and S4, iteratively calculating the optimal allocation result of each patient by using the Italian algorithm and the Newton-Raphson optimization algorithm, so that each patient can be first-aid. Through combination optimization of multiple algorithms, the method is superior to a traditional single-algorithm scheduling scheme in the aspects of response time, treatment effect and resource utilization rate, multi-ambulance task allocation can be efficiently carried out, and it is ensured that each patient can be effectively treated in time.

Description

Technical Field

[0001] The present invention relates to the field of intelligent scheduling and optimization technology, and in particular to a 120 emergency center ambulance task allocation method based on multi-algorithm combination optimization. Background Art

[0002] The existing ambulance dispatch system mainly relies on manual dispatch or simple rule-based algorithms, which makes it difficult to achieve efficient task allocation and path optimization in complex dynamic environments. Traditional artificial intelligence algorithms have great limitations in dealing with the randomness and suddenness of emergency tasks, resulting in long response times and low resource utilization, which cannot fully meet actual needs. Therefore, a new algorithm is urgently needed to improve the efficiency and response speed of ambulance dispatch.

[0003] Although the application of artificial intelligence algorithms in other fields has achieved remarkable results, their application in the 120 ambulance dispatch system still faces many challenges. Many existing artificial intelligence-based dispatch algorithms, such as genetic algorithms and ant colony algorithms, have outstanding performance in solving certain optimization problems. However, due to the particularity of ambulance dispatch problems, these algorithms have not been able to fully adapt to the highly dynamic, multivariable and real-time dispatch environment. For example, when dealing with large-scale dispatch problems, genetic algorithms are prone to fall into local optimality and are difficult to achieve global optimal solutions; although ant colony algorithms can quickly find feasible solutions in the early stage, they have defects in real-time path updates and optimization decisions. This shows that the existing single artificial intelligence algorithms have limitations in the field of ambulance dispatch, especially in dealing with real-time dispatch problems in complex and dynamic environments. Summary of the invention

[0004] In order to solve the problem that the existing single artificial intelligence algorithm cannot fully adapt to the complex emergency vehicle dispatching needs in actual situations, the present invention proposes a 120 emergency center ambulance task allocation method based on multi-algorithm combination optimization to solve the above problem.

[0005] The present application discloses a multi-algorithm combination optimization method for allocating ambulance tasks to a 120 emergency center, comprising the following steps:

[0006] S1. Obtain the specific locations of all patients waiting for treatment and their severity within the current time period, the specific locations of all idle ambulances, the specific locations of all hospitals, and the hospitals' ability to treat each patient.

[0007] S2. Construct an evaluation function for the ambulance's task allocation and an evaluation function for the treatment effect of each patient.

[0008] S3. Combine the A* algorithm for path planning to ensure that the ambulance chooses the optimal route.

[0009] S4. Use Ito algorithm and Newton-Raphson optimization algorithm to iteratively calculate the optimal allocation result for each patient so that each patient can receive emergency treatment.

[0010] Preferably, the S1 comprises the following steps:

[0011] The specific locations of all patients waiting for treatment in the current time period can be obtained through the emergency call system, which can be obtained through the GPS positioning system, emergency call records, and location services of the patient's mobile phone.

[0012] The patient's serious condition is obtained through the patient's self-report during the emergency call, the emergency personnel's initial judgment and the patient's historical health records.

[0013] The specific locations of all idle ambulances are obtained through the vehicle dispatch system and GPS positioning system. The real-time location information of each ambulance is obtained through the GPS device installed on the vehicle and transmitted to the dispatch center through the wireless network.

[0014] The specific location of the hospital was obtained through geographic information system.

[0015] The hospital's ability to treat patients is obtained through the number of beds, emergency room capacity, available medical equipment, and number of medical staff in the hospital's real-time resource management system.

[0016] Preferably, the task allocation evaluation function is:

[0017]

[0018] Where N is the number of patients, q i is the treatment effect of the i-th patient, σ F is the standard deviation of the treatment effect of all patients, and λ is an adjustment parameter used to balance the total treatment effect and its standard deviation. This formula indicates that in order to make the treatment effect of each patient relatively balanced, so as not to endanger the life of patients with poor treatment effect, the standard deviation of the treatment effect of all patients σ F Balance the task distribution of all ambulances.

[0019] Preferably, the treatment effect evaluation function is:

[0020]

[0021] Among them, E ij is the treatment capacity of the jth hospital for the ith patient, T i is the time it takes for the ambulance to reach the location of the i-th patient, T ij is the time consumed by the i-th patient from waiting for the ambulance to arriving at the j-th hospital, W iis the severity coefficient of the i-th patient.

[0022] First, according to the specific location of a patient and the specific location of an ambulance, the time consumption T of the ambulance going to the patient is obtained. i Then, based on the specific location of the patient and the specific location of a hospital, the time consumption T for the patient to go to the hospital is obtained. ij , then the time consumed by the patient from dialing 120 to arriving at a hospital is T i +T ij In order to get better treatment for patients, it is necessary to find the best ambulance and hospital. ij The maximum allocation result. This treatment effect evaluation function is used to evaluate the treatment effect of different hospital and ambulance combinations on patients, and is used to optimize the allocation results in subsequent steps.

[0023] Preferably, S3 comprises the following steps:

[0024] S31. Define a state space, where the state space includes all road network nodes, each node representing a geographical location, such as a street intersection, a hospital, a patient location, etc.

[0025] S32. Define a heuristic function, which is used to estimate the shortest path from the current node to the target node. The heuristic function is as follows:

[0026]

[0027] Among them, (x goal ,y goal ) is the coordinate of the target node, (x n ,y n ) is the coordinate of the current node;

[0028] S33, initialization, initializing the open list and the closed list, adding the starting point node to the open list, and setting its initial cost to 0.

[0029] S34, node expansion, select the node with the smallest cost function from the open list for expansion, the cost function is as follows:

[0030] f(n)=g(n)+h(n);

[0031] Among them, g(n) is the actual cost from the starting point to the current node.

[0032] S35, path update, for each adjacent node of the current node, calculate the new cost function value f(n), if the adjacent node is not in the open list, or the new cost function value is less than the existing value of the node, then update the cost value of the node and set its parent node as the current node.

[0033] S36. Repeat the iteration, repeat S34 and S35 until the target node is expanded, or the open list is empty (indicating that there is no feasible path).

[0034] S37, path backtracking, starting from the target node, and gradually backtracking to the starting node through the parent node pointer to obtain the optimal path.

[0035] Preferably, S4 comprises the following steps:

[0036] S41, particle fitness calculation, calculates the fitness of all particles according to the task evaluation function formula obtained in S2, and the value of the evaluation function F calculated is the particle fitness.

[0037] S42. Sort all particles according to their fitness, and calculate the radius of each particle according to the fitness of each particle. The calculation formula is:

[0038]

[0039] Among them, r max is the maximum particle radius, r min is the minimum value of the particle radius, p represents the particle ranked in the pth position by fitness, and M is the total number of particles.

[0040] S43. Calculate the ambient temperature to simulate the cooling process during the distribution process and avoid the local optimal solution. The calculation formula is:

[0041] τ t =τ0×α t ;

[0042] Among them, τ t is the ambient temperature at the tth iteration, τ0 is the initial temperature, and α is the annealing coefficient.

[0043] S44. Calculate the motion intensity of the particles. The calculation formula is:

[0044]

[0045] Among them, β is the influence factor of particle radius.

[0046] S45, executing the drift operator. The purpose of the drift operator is to move the solution closer to the global optimal solution, reduce the deviation of the solution, and ensure that each patient can get the best emergency treatment effect, which is expressed by the following formula:

[0047] H i (s,j)=H i (s,j)+X(p)σ″,ifE c (s,j)∈σ∪σ′;

[0048] Among them, σ is the current particle, σ′ is the optimal particle in this round of iteration, that is, the attractor, and σ″ represents the ambulance and hospital that have not been assigned to the i-th patient in this round of iteration, which is H i (s,j) represents the ambulance s and hospital j to which the i-th patient is assigned, E c (s,j) represents the edge of node (s,j) in the cth iteration. This formula indicates increasing the selection weight of the same allocation results in the current particle and the optimal particle, so that the feasible solution of the next iteration drifts toward the optimal solution.

[0049] S46. Execute the wave operator. The purpose of the wave operator is to increase the diversity of solutions and prevent falling into the local optimum. It is expressed by the following formula:

[0050] H i (s,j)=H i (s,j)+X(p),ifE c (s,j)∈σ″∪σ′∪σ″∩rand(0,1) <P perturb ;

[0051] Among them, rand(0,1) represents a random function that can generate a random number in the interval [0,1]. perturb is the perturbation probability, which is used to control the probability of the particle to undergo a fluctuation process. This formula indicates that the number of ambulances and hospitals that have not been assigned to a patient is increased, as well as the selection weight of a random ambulance and hospital, so that the particles in the next iteration are randomly perturbed.

[0052] S47. Update the particles and construct a new solution.

[0053] S48. If the updated new particle is better than the current particle, the current particle is replaced with the new particle and the update of the current particle is ended.

[0054] S49, determine whether the current optimal particle is better than the historical optimal particle. If so, replace the historical optimal particle with the current optimal particle and reset the maximum number of unupdated times to 0. If not, add 1 to the maximum number of unupdated times.

[0055] S410, determine whether the maximum number of unupdated times has been reached, if so, output the optimal allocation result, if not, go to S43.

[0056] S411. Combine the Newton-Raphson optimization algorithm to further optimize the allocation result of each patient.

[0057] Preferably, the S47 comprises the following steps:

[0058] S471, initialize the taboo table to be empty;

[0059] S472, traverse each patient to construct a new feasible solution;

[0060] S473, add the ambulances that are not in the taboo list to the candidate list;

[0061] S474. Calculate the selection probabilities of candidate ambulances and all hospitals. The calculation formula is as follows:

[0062]

[0063] Among them, P i (s, j) represents the probability of selecting the ambulance and hospital (s, j) assigned to the i-th patient, tabu represents the taboo table, which records all the assigned ambulances, η i (j) is T ij The reciprocal of , the parameter Ψ is used to control the weight of pheromone concentration in the selection probability calculation, and the parameter ω is used to control the weight of time consumption in the selection probability calculation;

[0064] S475, using roulette to select an ambulance and a hospital to be assigned to the current patient;

[0065] S476, remove the ambulance from the candidate list and add it to the taboo list;

[0066] S477. Determine whether all patients have been assigned. If so, output a new solution and end the current loop. Otherwise, return to S472.

[0067] Preferably, the S411 comprises the following steps:

[0068] S4111. Construct an objective function so that each patient can get the best first aid effect. The objective function calculation formula is as follows:

[0069]

[0070] Where x is the vector of all patient allocation plans;

[0071] S4112. Calculate the gradient of the objective function. The calculation formula is:

[0072]

[0073] where each component x1,x2,…,x N represents a decision variable;

[0074] S4113. Calculate the Hessian matrix of the objective function. The calculation formula is:

[0075]

[0076] Among them, xa and x b are the decision variables in the solution vector, representing different dimensions in the optimization problem. The elements of the Hessian matrix Represents the objective function G with respect to the variable x a and x b The second-order partial derivative of , which reflects the curvature information of the function in all directions.

[0077] 41114. Update the allocation scheme according to the Newton-Raphson optimization algorithm. The update formula is:

[0078]

[0079] Among them, x k is the allocation scheme for the kth iteration;

[0080] S4115. Set the iteration termination condition. When the norm of the gradient is less than the preset threshold ε, stop the iteration and output the optimal allocation solution.

[0081] Beneficial effects of the present invention:

[0082] (1 The present invention provides efficient path optimization through the A* algorithm, ensuring that the ambulance selects the optimal route and shortens the response time. The randomness and dynamic changes of emergency tasks are processed by the Ito algorithm, providing robustness and adaptability, and ensuring the flexibility of task allocation. The Newton-Raphson optimization algorithm provides fast local optimization capabilities, and combined with the initial solution of the Ito algorithm, the accuracy of the allocation results is further improved.

[0083] (2) Through the combined optimization of multiple algorithms, the present invention is superior to the traditional single-algorithm scheduling scheme in terms of response time, treatment effect and resource utilization, fully demonstrating its advantages in complex dynamic environments, and can efficiently allocate tasks to multiple ambulances to ensure that every patient can receive timely and effective treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Figure 1 This is a flow chart of the 120 emergency center ambulance task allocation method based on multi-algorithm combination optimization according to an embodiment of the present invention. DETAILED DESCRIPTION

[0085] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and examples.

[0086] The embodiment of the present application discloses a method for allocating ambulance tasks of a 120 emergency center based on a combination of multiple algorithms. The process is as follows: Figure 1 As shown, the following steps are included:

[0087] S1. Obtain the specific locations of all patients waiting for treatment and their severity within the current time period, the specific locations of all idle ambulances, the specific locations of all hospitals, and the hospitals' ability to treat each patient.

[0088] The specific locations of all patients waiting for treatment in the current time period can be obtained through the emergency call system, which can be obtained through the GPS positioning system, emergency call records, and location services of the patient's mobile phone.

[0089] In a specific embodiment, the patient's serious condition is obtained through the patient's self-report during the emergency call, the initial judgment of the emergency personnel, and the patient's historical health record. The specific location of all idle ambulances is obtained through the vehicle dispatch system and GPS positioning system. The real-time location information of each ambulance is obtained through the GPS device installed on the vehicle and transmitted to the dispatch center through the wireless network. The specific location of the hospital is obtained through the geographic information system. The hospital's ability to treat patients is obtained through the number of beds, emergency room capacity, available medical equipment, and the number of medical staff in the hospital's real-time resource management system.

[0090] S2. Construct an evaluation function for the ambulance's task allocation and an evaluation function for the treatment effect of each patient.

[0091] The calculation formula of the task allocation evaluation function is:

[0092]

[0093] Where N is the number of patients, q i is the treatment effect of the i-th patient, σ F is the standard deviation of the treatment effect of all patients, and λ is an adjustment parameter used to balance the total treatment effect and its standard deviation. The task allocation evaluation function indicates that in order to make the treatment effect of each patient relatively balanced, so as not to endanger the life of patients with poor treatment effect, the standard deviation of the treatment effect of all patients σ F Balance the task distribution of all ambulances.

[0094] The calculation formula of the treatment effect evaluation function is:

[0095]

[0096] Among them, E ij is the treatment capacity of the jth hospital for the ith patient, T i is the time it takes for the ambulance to reach the location of the i-th patient, T ij is the time consumed by the i-th patient from waiting for the ambulance to arriving at the j-th hospital, W i is the severity coefficient of the i-th patient.

[0097] First, according to the specific location of a patient and the specific location of an ambulance, the time consumption T of the ambulance going to the patient is obtained. i Then, based on the specific location of the patient and the specific location of a hospital, the time consumption T for the patient to go to the hospital is obtained. ij , then the time consumed by the patient from dialing 120 to arriving at a hospital is T i +T ij In order to get better treatment for patients, it is necessary to find the best ambulance and hospital. ij The maximum allocation result. This treatment effect evaluation function is used to evaluate the treatment effect of different hospital and ambulance combinations on patients, and is used to optimize the allocation results in subsequent steps.

[0098] S3. Combine the A* algorithm for path planning to ensure that the ambulance chooses the optimal route.

[0099] S31. Define a state space, which includes all road network nodes. Each node represents a geographical location, such as a street intersection, a hospital, a patient location, etc.

[0100] S32. Define a heuristic function, which is used to estimate the shortest path from the current node to the target node. The heuristic function is as follows:

[0101]

[0102] Among them, (x goal ,y goal ) is the coordinate of the target node, (x n ,y n ) are the coordinates of the current node.

[0103] S33, initialization, initializing the open list and the closed list, adding the starting point node to the open list, and setting its initial cost to 0.

[0104] S34, node expansion, select the node with the smallest cost function from the open list for expansion, the cost function is as follows:

[0105] f(n)=g(n)+h(n);

[0106] Among them, g(n) is the actual cost from the starting point to the current node.

[0107] S35, path update, for each adjacent node of the current node, calculate the new cost function value f(n), if the adjacent node is not in the open list, or the new cost function value is less than the existing value of the node, then update the cost value of the node and set its parent node as the current node.

[0108] S36. Repeat the iteration, repeat S34 and S35 until the target node is expanded, or the open list is empty (indicating that there is no feasible path).

[0109] S37, path backtracking, starting from the target node, and gradually backtracking to the starting node through the parent node pointer to obtain the optimal path.

[0110] S4. Use Ito algorithm and Newton-Raphson optimization algorithm to iteratively calculate the optimal allocation result for each patient so that each patient can receive emergency treatment.

[0111] S41, particle fitness calculation, calculates the fitness of all particles according to the task evaluation function formula obtained in S2, and the value of the evaluation function F calculated is the particle fitness.

[0112] S42. Sort all particles according to their fitness, and calculate the radius of each particle according to the fitness of each particle. The calculation formula is:

[0113]

[0114] Among them, r max is the maximum value of the particle radius, which is set to 1 in this embodiment. min is the minimum value of the particle radius, which is set to 0 in this embodiment, p represents the particle ranked at the pth position according to the fitness size, and M is the total number of particles.

[0115] S43. Calculate the ambient temperature to simulate the cooling process during the distribution process and avoid the local optimal solution. The calculation formula is:

[0116] τ t =τ0×α t ;

[0117] Among them, τ y is the ambient temperature at the tth iteration, τ0 is the initial temperature, and α is the annealing coefficient, which is set to 0.9 in this embodiment.

[0118] S44. Calculate the motion intensity of the particles. The calculation formula is:

[0119]

[0120] Wherein, β is the influence factor of particle radius, which is set to 2 in this embodiment.

[0121] S45, executing the drift operator. The purpose of the drift operator is to move the solution closer to the global optimal solution, reduce the deviation of the solution, and ensure that each patient can get the best emergency treatment effect, which is expressed by the following formula:

[0122] Hi (s,j)=H i (s,j)+X(p)σ″,ifE c (s,j)∈σ∪σ′;

[0123] Among them, σ is the current particle, σ ′ is the optimal particle in this round of iteration (the particle with the highest fitness F is the optimal particle), that is, the attractor, σ″ represents the ambulance and hospital that have not been assigned to the i-th patient in this round of iteration, which is H i (s,j) represents the ambulance s and hospital j to which the i-th patient is assigned, E c (s,j) represents the edge of node (s,j) in the cth iteration. This formula indicates increasing the selection weight of the same allocation results in the current particle and the optimal particle, so that the feasible solution of the next iteration drifts toward the optimal solution.

[0124] S46. Execute the wave operator. The purpose of the wave operator is to increase the diversity of solutions and prevent falling into the local optimum. It is expressed by the following formula:

[0125] H i (s,j)=H i (s,j)+X(p),ifE c (s,j)∈σ″∪σ′∪σ″∩rand(0,1) <P perturb ;

[0126] Among them, rand(0,1) represents a random function that can generate a random number in the interval [0,1]. perturb is the perturbation probability, which is used to control the probability of the particle to undergo a fluctuation process. This formula indicates that the number of ambulances and hospitals that have not been assigned to a patient is increased, as well as the selection weight of a random ambulance and hospital, so that the particles in the next iteration are randomly perturbed.

[0127] S47. Update the particles and construct a new solution.

[0128] S471. Initialize the taboo table to be empty.

[0129] S472. Traverse each patient and construct a new feasible solution.

[0130] S473. Add the ambulance that is not in the taboo list to the candidate list.

[0131] S474. Calculate the selection probabilities of candidate ambulances and all hospitals. The calculation formula is as follows:

[0132]

[0133] Among them, P i(s, j) represents the probability of selecting the ambulance and hospital (s, j) assigned to the i-th patient, tabu represents the taboo table, which records all the assigned ambulances, η i (j) is T ij The reciprocal of , the parameter ψ is used to control the weight of pheromone concentration (the selection weight of ambulance and hospital) in the selection probability calculation, and the parameter ω is used to control the weight of time consumption in the selection probability calculation.

[0134] S475. Use roulette to select an ambulance and a hospital to be assigned to the current patient.

[0135] S476. Remove the ambulance from the candidate list and add it to the taboo list.

[0136] S477. Determine whether all patients have been assigned. If so, output a new solution and end the current loop. Otherwise, return to S472.

[0137] S48, if the updated new particle is better than the current particle, the new particle is used to replace the current particle and the current particle update is terminated. The larger the value F of the fitness function is, the better the solution is. By comparing the fitness function, it is ensured that the particles gradually move toward the global optimal solution in each iteration.

[0138] S49, determine whether the current optimal particle is better than the historical optimal particle. If so, replace the historical optimal particle with the current optimal particle and reset the maximum number of unupdated times to 0. If not, add 1 to the maximum number of unupdated times.

[0139] S410, determine whether the maximum number of unupdated times has been reached. In this implementation, the maximum number of unupdated times is 20. If yes, output the optimal allocation result. If no, go to S43;

[0140] S411. Combine the Newton-Raphson optimization algorithm to further optimize the allocation result of each patient.

[0141] S4111. Construct an objective function so that each patient can get the best first aid effect. The objective function calculation formula is as follows:

[0142]

[0143] Where x is the vector of all patient allocation plans;

[0144] S4112. Calculate the gradient of the objective function. The calculation formula is:

[0145]

[0146] Here, each component represents a decision variable in the problem.

[0147] S4113. Calculate the Hessian matrix of the objective function. The calculation formula is:

[0148]

[0149] Among them, x a and x b are the decision variables in the solution vector, representing different dimensions in the optimization problem. The elements of the Hessian matrix Represents the objective function G with respect to the variable x a and x b The second-order partial derivative of , which reflects the curvature information of the function in all directions.

[0150] 41114. Update the allocation scheme according to the Newton-Raphson optimization algorithm. The update formula is:

[0151]

[0152] Among them, x k is the allocation scheme for the kth iteration;

[0153] S4115, set the iteration termination condition. When the norm of the gradient is less than the preset threshold ε, stop the iteration and output the optimal allocation solution. In this embodiment, ε=10 -5 .

[0154] In a specific embodiment, in order to evaluate the effect of the embodiment of the present application in practical application, a simulation experiment is set up, and compared with the Floyd algorithm, the ant colony algorithm and the single algorithm included in the embodiment of the present application. The specific scheme of the simulation is as follows:

[0155] The historical data of a city’s 120 emergency center is selected, including the patient’s location, ambulance location, hospital location and its rescue capacity.

[0156] The average response time, average treatment effect and resource utilization were selected as evaluation indicators.

[0157] The average response time refers to the average time from receiving an emergency call to the ambulance arriving at the patient's location. The calculation formula is:

[0158]

[0159] The average treatment effect refers to the average of the treatment effects obtained by all patients, and the calculation formula is:

[0160]

[0161] Resource utilization refers to the efficiency of the use of ambulance and hospital resources. It is used to measure the utilization of resources by the dispatch system. The calculation formula is:

[0162]

[0163] in, Indicates the number of all used ambulances, that is, the number of ambulance resources used. represents the number of all hospitals used, that is, the number of hospital resources used, S represents the total number of ambulances, and J represents the total number of hospitals.

[0164] The road network model selects the actual road network data of a city, including main roads and intersections. The default vehicle speed is 60km / h. Two simulation scenarios are set. In scenario 1, all the patients who are called have the same treatment weights. The simulation test results are shown in Table 1. In scenario 2, different treatment weights are set according to the actual condition. The simulation test results are shown in Table 2:

[0165] Table 1 Evaluation index results of each method when the rescue weight is the same

[0166]

[0167] Table 2 Evaluation index results of each method with different rescue weights

[0168]

[0169]

[0170] Tables 1 and 2 show the experimental results in different simulation scenarios. By comparing with the Floyd algorithm and the ant colony algorithm, it can be seen that the method of the present application is superior in response time, treatment effect and resource utilization, and this advantage is more obvious in the environment of giving treatment weight.

[0171] Table 3 shows the comparison between the combined algorithm of this application and the single algorithms included. The combined optimization algorithm is superior to the single algorithm in terms of average response time, average rescue effect and resource utilization, which verifies the superiority of the combined algorithm in improving overall performance.

[0172] Table 3 Evaluation index results of the combined algorithm and each single algorithm of this application

[0173]

[0174] Through the above comprehensive evaluation index comparison, it can be seen that in both scenarios, the method of the present application significantly shortens the response time, ensuring that patients can receive treatment more quickly; it can effectively improve the average treatment effect of patients, and through the evaluation function σ FIt plays a role in ensuring the balance of treatment effects and avoiding the situation where the treatment effect is too poor; at the same time, it also performs well in terms of resource utilization, and can use existing ambulances and medical resources more efficiently while maintaining a certain redundancy. It can be seen that the method of the present application is more efficient in emergency vehicle dispatch, especially when the treatment weight (patient condition and hospital treatment level) is taken into account. This advantage is more obvious. It fully demonstrates the advantages of the present application in ambulance task allocation.

[0175] In summary, the 120 emergency center ambulance task allocation method based on multi-algorithm combination optimization provided in this application can significantly improve the dispatch efficiency and treatment effect in practical applications and realize efficient utilization of resources.

[0176] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. The multi-algorithm combination optimization method for allocating tasks of ambulances in the 120 emergency center is characterized by: The following steps are involved: S1. Obtain the specific location and severity of all patients waiting for treatment in the current time period, the specific location of all idle ambulances, the specific location of all hospitals and the hospital's ability to treat each patient; S2, constructing an evaluation function for the task allocation of the ambulance and an evaluation function for the treatment effect of each patient; S3, combine A* algorithm to perform path planning to ensure that the ambulance chooses the optimal route; S4. Use Ito algorithm and Newton-Raphson optimization algorithm to iteratively calculate the optimal allocation result for each patient so that each patient can receive emergency treatment.

2. The 120 emergency center ambulance task allocation method based on multi-algorithm combination optimization according to claim 1 is characterized in that: The S1 comprises the following steps: Obtain the specific locations of all patients waiting for treatment in the current time period through the emergency call system; The patient's serious condition is obtained through the patient's self-report during the emergency call, the emergency personnel's initial judgment, and the patient's historical health records; Obtain the specific locations of all available ambulances through the vehicle dispatch system and GPS positioning system; Obtain the specific location of the hospital through geographic information system; Obtain the hospital's ability to treat patients through the hospital's real-time resource management system.

3. The 120 emergency center ambulance task allocation method based on multi-algorithm combination optimization according to claim 2 is characterized in that: The task allocation evaluation function is: Where N is the number of patients, q i is the treatment effect of the i-th patient, σ F is the standard deviation of the treatment effect of all patients, and λ is the adjustment parameter.

4. The 120 emergency center ambulance task allocation method based on multi-algorithm combination optimization according to claim 3 is characterized in that: The treatment effect evaluation function is: Among them, E ij is the treatment capacity of the jth hospital for the ith patient, T i is the time it takes for the ambulance to reach the location of the i-th patient, T ij is the time consumed by the i-th patient from waiting for the ambulance to arriving at the j-th hospital, W i is the severity coefficient of the i-th patient.

5. The 120 emergency center ambulance task allocation method based on multi-algorithm combination optimization according to claim 4 is characterized in that: The S3 comprises the following steps: S31, defining a state space, wherein the state space includes all road network nodes, and each node represents a geographical location; S32. Define a heuristic function, which is used to estimate the shortest path from the current node to the target node. The heuristic function is as follows: Among them, (x goal ,y goal ) is the coordinate of the target node, (x n ,y n ) is the coordinate of the current node; S33, initialization, initializing the open list and the closed list, adding the starting point node to the open list, and setting its initial cost to 0; S34, node expansion, select the node with the smallest cost function from the open list for expansion, the cost function is as follows: f(n)=g(n)+h(n); Among them, g(n) is the actual cost from the starting point to the current node; S35, path update, for each adjacent node of the current node, calculate a new cost function value f(n), if the adjacent node is not in the open list, or the new cost function value is less than the existing value of the node, then update the cost value of the node and set its parent node as the current node; S36, repeat iteration, repeat S34 and S35 until the target node is expanded or the open list is empty; S37, path backtracking, starting from the target node, and gradually backtracking to the starting node through the parent node pointer to obtain the optimal path.

6. The multi-algorithm combination optimization 120 emergency center ambulance task allocation method according to claim 5 is characterized in that: The S4 comprises the following steps: S41, particle fitness calculation, calculate the fitness of all particles according to the task evaluation function formula obtained in S2, and the value of the evaluation function F calculated is the particle fitness; S42. Sort all particles according to their fitness, and calculate the radius of each particle according to the fitness of each particle. The calculation formula is: Among them, r max is the maximum particle radius, r min is the minimum value of the particle radius, p represents the particle ranked at the pth position according to the fitness size, and M is the total number of particles; S43. Calculate the ambient temperature to simulate the cooling process during the distribution process and avoid the local optimal solution. The calculation formula is: t t =τ0×α t ; Among them, τ t is the ambient temperature at the tth iteration, τ0 is the initial temperature, and α is the annealing coefficient; S44. Calculate the motion intensity of the particles. The calculation formula is: Among them, β is the influencing factor of particle radius; S45, executing the drift operator. The purpose of the drift operator is to move the solution closer to the global optimal solution, reduce the deviation of the solution, and ensure that each patient can get the best emergency treatment effect, which is expressed by the following formula: H i (s,j)=H i (s,j)+X(p)σ″,if E c (s,j)∈σ∪σ′; Where σ is the current particle, σ′ is the optimal particle in this iteration, i.e., the attractor, σ″ represents the ambulance and hospital that have not been assigned to the i-th patient in this iteration, Hi(s, j) represents the ambulance s and hospital j to which the i-th patient is assigned, and E c (s, j) represents the edge of node (s, j) in the cth iteration; S46. Execute the wave operator. The purpose of the wave operator is to increase the diversity of solutions and prevent falling into the local optimum. It is expressed by the following formula: H i (s,j)=H i (s,j)+X(p),if E c (s,j)∈σ″∪σ′∪σ″∩rand(0,1)<P perturb ; Among them, rand(0,1) represents a random function that can generate a random number in the interval [0,1]. perturb is the disturbance probability; S47, update the particles and construct a new solution; S48, if the updated new particle is better than the current particle, replace the current particle with the new particle and end the current particle update; S49, judging whether the current optimal particle is better than the historical optimal particle, if so, replacing the historical optimal particle with the current optimal particle, and resetting the maximum number of unupdated times to 0, if not, adding 1 to the maximum number of unupdated times; S410, determine whether the maximum number of unupdated times has been reached, if so, output the optimal allocation result, if not, go to S43; S411. Combine the Newton-Raphson optimization algorithm to further optimize the allocation result of each patient.

7. The multi-algorithm combination optimization 120 emergency center ambulance task allocation method according to claim 6 is characterized in that: The S47 comprises the following steps: S471, initialize the taboo table to be empty; S472, traverse each patient to construct a new feasible solution; S473, add the ambulances that are not in the taboo list to the candidate list; S474. Calculate the selection probabilities of candidate ambulances and all hospitals. The calculation formula is as follows: Where Pi(s, j) represents the probability of selecting the ambulance and hospital (s, j) assigned to the i-th patient, tabu represents the taboo table, which records all the assigned ambulances, and η i (j) is T ij The reciprocal of , the parameter Ψ is used to control the weight of pheromone concentration in the selection probability calculation, and the parameter ω is used to control the weight of time consumption in the selection probability calculation; S475, using roulette to select an ambulance and a hospital to be assigned to the current patient; S476, remove the ambulance from the candidate list and add it to the taboo list; S477. Determine whether all patients have been assigned. If so, output a new solution and end the current loop. Otherwise, return to S472.

8. The 120 emergency center ambulance task allocation method based on multi-algorithm combination optimization according to claim 7 is characterized in that: The S411 includes the following steps: S4111. Construct an objective function so that each patient can get the best first aid effect. The objective function calculation formula is as follows: Where x is the vector of all patient allocation plans; S4112. Calculate the gradient of the objective function. The calculation formula is: where each component x1, x2, ..., x N represents a decision variable; S4113. Calculate the Hessian matrix of the objective function. The calculation formula is: Among them, x a and x b is the decision variable in the solution vector, representing the different dimensions in the optimization problem, and the elements of the Hessian matrix Represents the objective function G with respect to the variable x a and x b The second-order partial derivative of ; 41114. Update the allocation scheme according to the Newton-Raphson optimization algorithm. The update formula is: Among them, x k is the allocation scheme for the kth iteration; S4115. Set the iteration termination condition. When the norm of the gradient is less than the preset threshold ε, stop the iteration and output the optimal allocation solution.

Images