Genetic algorithm-based travel itinerary planning system

Abstract

The invention discloses a travel itinerary planning system based on a genetic algorithm. The travel itinerary planning system comprises a big data module, an initialization module, an optimization iteration module and a visual feedback module, basic information of each scenic spot is stored in the big data module, and the initialization module receives a travel itinerary planning request from a traveler based on a genetic algorithm, extracts scenic spot information conforming to the planning request from the big data module, pre-processes and classifies the scenic spot information to generate an initial population, and sends the initial population to the big data module; the optimization iteration module migrates the variation population to a new population based on reinforcement learning of transfer learning, and introduces a reinforcement learning mechanism, and the visual feedback module receives a real-time optimized travel route obtained by the optimization iteration module, and outputs an optimal travel itinerary plan in a visual form, so that the optimal travel itinerary plan can be obtained in use. According to requirements and preferences of tourists, personalized tourism routes can be planned, so that tourism experience and satisfaction of the tourists are improved.

Description

Technical Field

[0001] The invention relates to a travel route planning system, in particular to a travel itinerary planning system based on genetic algorithm. Background Art

[0002] Genetic algorithm is an optimization algorithm that simulates the natural evolution process. It simulates the operations of selection, crossover and mutation in the biological genetic evolution process, and optimizes candidate solutions through continuous evolution to find the optimal solution. In travel path planning, genetic algorithm can help find the optimal path to minimize the overall travel distance.

[0003] Travel route planning is a classic combinatorial optimization problem with a wide range of applications. Suppose a traveling salesman wants to travel between multiple cities. Our goal is to find an optimal path that minimizes the overall distance of the trip. This problem is a typical traveling salesman problem and cannot be accurately solved by traditional computational methods. As a heuristic algorithm, genetic algorithms have strong global search capabilities and can find better solutions within an acceptable time.

[0004] However, existing travel itinerary planning systems based on genetic algorithms are often based on fixed templates or popular routes, lacking personalized customization for tourists' personal needs and preferences; and although genetic algorithms can simulate the search for optimal solutions in the natural evolution process, their adaptability and flexibility are still limited when faced with complex and changeable practical problems. Therefore, genetic algorithms can easily fall into local optimal solutions during the search process, thus affecting the quality of travel itinerary planning.

[0005] Therefore, there is an urgent need for a travel itinerary planning system based on genetic algorithms to solve the technical problems existing in the above-mentioned prior art. Summary of the invention

[0006] The invention overcomes the shortcomings of the prior art and provides a travel itinerary planning system based on genetic algorithms.

[0007] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is: a travel itinerary planning system based on genetic algorithm, comprising: a big data module, an initialization module, an optimization iteration module and a visual feedback module;

[0008] The big data module stores basic information of each scenic spot, including the name, geographical location, scenic grade, opening hours, ticket price, previous passenger flow, and accommodation conditions of the scenic spot;

[0009] The initialization module receives a travel itinerary planning request from a traveler based on a genetic algorithm, extracts scenic spot information that meets the planning request from the big data module, pre-processes and classifies the scenic spot information, generates an initial population, establishes a fitness function in the initialization module, and is used to calculate and evaluate the quality of each individual in the initial population; selects excellent individuals according to the fitness value, crosses some genes of the excellent individuals to generate new individuals, and randomly mutates the new individuals to generate a mutant population;

[0010] The optimization iteration module migrates the variant population to a new population based on the reinforcement learning of transfer learning, and introduces a reinforcement learning mechanism to optimize the travel route in real time after multiple iterative calculations based on the real-time feedback of travelers;

[0011] The visualization feedback module receives the real-time optimized travel route obtained from the optimization iteration module, and outputs the optimal travel itinerary planning in a visualized form.

[0012] In a preferred embodiment of the present invention, the process of establishing the initial population includes: obtaining the name and geographical location information of the scenic spot, preprocessing the geographical location information, converting the coordinate information of the geographical location into matrix information using standard Euclidean distance, and generating population individuals containing multiple travel routes.

[0013] In a preferred embodiment of the present invention, the initialization module establishes a fitness function by considering the scenic grade, opening hours, ticket price, previous passenger flow and accommodation conditions of the scenic spot, and calculates the quality of the individuals in the population. The formula expression of the fitness function is:

[0014] F(i)=α*S(i)+β*T(i)-γ*P(i)-δ*C(i)+ε*A(i)

[0015] Where F(i) is the fitness value of the population individual, which is used to evaluate the quality of the population individual; α, β, γ, δ, and ε are the weight coefficients of scenery grade, opening hours, ticket price, previous passenger flow, and accommodation conditions, respectively; S(i) is the scenery grade score of the i-th population individual in the initial population; T(i) is the convenience score of the opening hours of the i-th population individual in the initial population; P(i) is the score of the scenic spot ticket price of the i-th population individual in the initial population; C(i) is the score of the previous passenger flow of the i-th population individual in the initial population; A(i) is the score of the accommodation conditions of the i-th population individual in the initial population.

[0016] In a preferred embodiment of the present invention, the reinforcement learning process based on transfer learning includes: using transfer learning technology to extract the computational experience of the genetic algorithm from the mutant population, and transferring it into the reinforcement learning, establishing a reinforcement learning model based on the computational experience, training and optimizing the reinforcement learning model through a fine-tuning method, and establishing a reward mechanism function in the reinforcement learning model, wherein the reward mechanism function includes a state value function and an action value function, which are used to represent the expected reward of the state-action; combining the real-time feedback of the traveler, performing multiple iterative calculations, and each iterative calculation updates the strategy according to the difference and reward of the previous time to obtain a real-time optimized travel route.

[0017] In a preferred embodiment of the present invention, the formula calculation expression of the state value function is:

[0018]

[0019] The formula calculation expression of the action value function is:

[0020]

[0021] Where V Φ (s) represents the expected return obtained by following the strategy Φ starting from state s; Q Φ (s,a) represents the expected return after starting from state s, executing action a, and then following strategy Φ; s t and a t are the state and action at time t, r(s t ,a t ) is in state s t Execute action a t The immediate reward obtained after Φ is the expectation under strategy Φ; η t The discount rate used to calculate the present value of future rewards, between 0 and 1.

[0022] In a preferred embodiment of the present invention, the formula calculation expression for the multiple iterative calculations is:

[0023] VΦ(s)=E Φ [r(s,a)+ηVΦ(s′)s]

[0024] QΦ(s,a)=E Φ [r(s,a)+ηQΦ(s′,a′)s,a]

[0025] Where s′ is the new state after state s executes action a, that is, the result after each iteration; a′ is the action selected according to strategy Φ in state s′; the meanings of the other variables are the same as above.

[0026] In a preferred embodiment of the present invention, after multiple iterations, the optimal value formula of the state value function is calculated as follows:

[0027]

[0028] Where V * (s) is the state value function with the largest value among all strategies Φ; Q * (s,a) is the action-value function with the largest value among all strategies Φ; the meanings of the remaining variables are the same as above.

[0029] In a preferred embodiment of the present invention, a computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned system when executing the computer program.

[0030] In a preferred embodiment of the present invention, a computer-readable storage medium stores a computer program, which implements the steps of the above-mentioned system when executed by a processor.

[0031] The present invention solves the defects existing in the background technology and has the following beneficial effects:

[0032] (1) The genetic algorithm-based travel itinerary planning system provided by the present invention combines genetic algorithms, transfer learning and reinforcement learning. When used, it can plan personalized travel routes according to the needs and preferences of tourists, thereby improving the tourists' travel experience and satisfaction; and by optimizing the travel routes, it reduces travel costs, improves travel cost-effectiveness, avoids excessive use and waste of tourism resources, and promotes the sustainable development of the tourism industry.

[0033] (2) The present invention adopts the method of transfer learning to effectively transfer the computational experience, optimization strategy and feature representation of the genetic algorithm to the reinforcement learning technology, thereby avoiding the blind exploration of reinforcement learning, accelerating the learning process, improving the optimization effect, and enhancing the robustness and adaptability of the system. At the same time, transfer learning also promotes the algorithm integration and innovation between genetic algorithms and reinforcement learning, and provides new ideas and methods for technological progress in the field of tourism itinerary planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

[0035] Figure 1 is a system flow chart of a preferred embodiment of the present invention;

[0036] Figure 2 is a flowchart of an iterative algorithm implementation of a preferred embodiment of the present invention;

[0037] Figure 3 It is a curve change diagram of the iteration process of the preferred embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0040] To further explain here, reinforcement learning is a method of learning through interaction between an agent / robot / agent and an environment. In reinforcement learning, the agent selects actions based on the state of the environment and adjusts its behavior strategy by observing the feedback (reward or punishment) from the environment to achieve the goal of maximizing long-term cumulative rewards.

[0041] Transfer learning is a method that uses the learning experience of one task to accelerate the learning of another related task. In transfer learning, the model can adapt to new tasks more quickly by learning the feature representation of one task, especially when the new task has less data or the new task has a certain relevance to the original task.

[0042] like Figure 1 and Figure 2 As shown, a travel itinerary planning system based on genetic algorithm includes: a big data module, an initialization module, an optimization iteration module and a visualization feedback module.

[0043] The big data module stores basic information about each scenic spot, including its name, geographical location, scenic rating, opening hours, ticket price, past visitor flow, and accommodation conditions.

[0044] Furthermore, the big data module may use one or more of MySQL, Oracle, PostgreSQL, Hadoop HDFS, or column family storage databases.

[0045] The initialization module is based on a genetic algorithm. It receives travel itinerary planning requests from travelers, extracts scenic spot information that meets the planning requests from the big data module, preprocesses and classifies the scenic spot information, and generates an initial population. The process of establishing the initial population includes: obtaining the name and geographic location information of the scenic spot, preprocessing the geographic location information, using the standard Euclidean distance to convert the coordinate information of the geographic location into matrix information, and generating a population of individuals containing multiple itinerary routes.

[0046] A fitness function is established in the initialization module to calculate and evaluate the quality of each individual in the initial population. The initialization module establishes a fitness function by considering the scenic level, opening hours, ticket prices, previous passenger flow, and accommodation conditions of the scenic spot to calculate the quality of the individuals in the population. The formula of the fitness function is:

[0047] F(i)=α*S(i)+β*T(i)-γ*P(i)-δ*C(i)+ε*A(i)

[0048] Where F(i) is the fitness value of the population individual, which is used to evaluate the quality of the population individual; α, β, γ, δ, and ε are the weight coefficients of scenery grade, opening hours, ticket price, previous passenger flow, and accommodation conditions, respectively; S(i) is the scenery grade score of the i-th population individual in the initial population; T(i) is the convenience score of the opening hours of the i-th population individual in the initial population; P(i) is the score of the scenic spot ticket price of the i-th population individual in the initial population; C(i) is the score of the previous passenger flow of the i-th population individual in the initial population; A(i) is the score of the accommodation conditions of the i-th population individual in the initial population.

[0049] What needs to be further explained here is that the sum of the weight coefficients of scenery level α, opening hours β, ticket price γ, previous passenger flow δ, and accommodation conditions ε is 1, and the proportion of different factors in the population is determined by the traveler when planning a single travel itinerary.

[0050] Select excellent individuals according to their fitness values, cross some of their genes to generate new individuals, and randomly mutate the new individuals to generate a mutant population;

[0051] The optimization iteration module is based on reinforcement learning of transfer learning, which migrates the mutant population to the new population and introduces the reinforcement learning mechanism. According to the real-time feedback of travelers, after multiple iterative calculations, the travel route is optimized in real time;

[0052] Furthermore, in order to ensure the best optimization strategy for the optimization iteration module, the initial population is obtained from historical data through transfer learning to improve the initial quality of the population. The initial population is generated by combining heuristic algorithms (such as greedy algorithms and nearest neighbor algorithms) to ensure that the initial population contains high-quality solutions.

[0053] Multi-population strategy: Using a multi-population genetic algorithm, each population evolves independently, and the global search capability is improved through information exchange between populations

[0054] The reinforcement learning process based on transfer learning includes: using transfer learning technology to extract the computational experience of the genetic algorithm from the mutant population, and transfer it into the reinforcement learning, establishing a reinforcement learning model based on the computational experience, training and optimizing the reinforcement learning model through fine-tuning methods, and establishing a reward mechanism function in the reinforcement learning model. The reward mechanism function includes a state value function and an action value function, which are used to represent the expected reward of the state-action; the formula calculation expression of the state value function is:

[0055]

[0056] The formula calculation expression of the action value function is:

[0057]

[0058] Where V Φ (s) represents the expected return obtained by following the strategy Φ starting from state s; Q Φ (s,a) represents the expected return after starting from state s, executing action a, and then following strategy Φ; s t and a t are the state and action at time t, r(s t ,a t ) is in state s t Execute action a t The immediate reward obtained after Φ is the expectation under strategy Φ; η t The discount rate used to calculate the present value of future rewards, between 0 and 1.

[0059] Through transfer learning, the system can use the mutant population that has been optimized by the genetic algorithm as a starting point to quickly adapt and optimize the reinforcement learning model, significantly reducing the training time of the reinforcement learning stage.

[0060] Combined with the real-time feedback from travelers, multiple iterative calculations are performed. The formula for multiple iterative calculations is as follows:

[0061]

[0062] QΦ(s,a)=EΦ [r(s,a)+ηQΦ(s′,a′)s,a]

[0063] Where s′ is the new state after state s executes action a, that is, the result after each iteration; a′ is the action selected according to strategy Φ in state s′; the meanings of the other variables are the same as above.

[0064] Each iterative calculation updates the strategy based on the difference and reward of the previous calculation to obtain a real-time optimized travel route. After multiple iterative calculations, the optimal value formula of the state value function is calculated as follows:

[0065]

[0066] Where V * (s) is the state value function with the largest value among all strategies Φ; Q * (s,a) is the action-value function with the largest value among all strategies Φ; the other variables have the same meanings as above.

[0067] To further explain, the reinforcement learning process based on transfer learning also includes:

[0068] Knowledge transfer: The system uses transfer learning technology to extract useful knowledge and experience from existing travel itinerary planning cases or related fields, including successful travel routes, tourists' preference patterns, and the popularity of specific attractions;

[0069] Model initialization: Based on the transferred knowledge, the system initializes the reinforcement learning model, which includes setting the initial parameters of the model, defining the state space, action space, and reward function.

[0070] State definition: In reinforcement learning, the state usually represents the current environment or situation. In this system, the state can be defined as the current location of the travel route, the attractions visited, the remaining time or budget, etc.

[0071] Action selection: An action is an action that a reinforcement learning system can take. In travel itinerary planning, actions include selecting the next attraction to visit, adjusting the tour time, and changing accommodation arrangements.

[0072] Reward function: The reward function is used to evaluate the goodness of the actions taken by the agent. In this system, the reward is calculated based on multiple factors, such as the number of attractions visited, the sum of the scenery levels, and the time or money saved.

[0073] Policy Optimization: By continuously interacting with the environment, the agent learns and optimizes its policy, which defines the best actions to take in different states.

[0074] Real-time feedback and adjustment: During the iteration process, the system dynamically adjusts the travel route based on real-time feedback (such as tourist satisfaction, road condition changes, weather forecasts, etc.), thereby ensuring that the planned route is always consistent with the actual situation and providing the best travel experience.

[0075] Multiple Iterations: The reinforcement learning process usually involves multiple iterations, each of which updates the policy based on the feedback and rewards from the previous iteration.

[0076] Convergence judgment: The system continues to iterate until a certain convergence condition is met, such as the reward function reaches its maximum value, the strategy no longer changes significantly, or the preset number of iterations is reached.

[0077] The visualization feedback module receives the real-time optimized travel route obtained from the optimization iteration module and outputs the optimal travel itinerary planning in a visual form.

[0078] like Figure 3 As shown in the figure, the horizontal axis represents the number of iterations, which is a dimensionless quantity, indicating the number of rounds or repeated executions of the algorithm; the vertical axis represents the path length, which is also assumed to be a dimensionless quantity and is used to represent relative distances or distance ratios. From this figure, it can be seen that the path length drops sharply from 0 to about 100 iterations, indicating that the algorithm has been efficiently optimized in the early stage and quickly shortened the path length; from about 100 iterations to 300, the path length continues to decrease but at a slower rate, which reflects that the algorithm has entered a more detailed adjustment stage, further optimizing the path through gradual fine-tuning; since 300 iterations, the path length has stabilized and has almost no significant changes, indicating that the algorithm has approached the optimal solution and further iterations have little effect on improving the path length.

[0079] When the present invention is used, the big data module serves as the information basis of the system and stores basic information of each scenic spot, including the name of the scenic spot, geographical location, scenery level, opening hours, ticket price, previous passenger flow and accommodation conditions.

[0080] Initialization module: Receives travel itinerary planning requests from travelers, clarifies planning objectives and constraints, extracts scenic spot information that meets the requirements from the big data module based on the planning request, and preprocesses and classifies the information, including obtaining the name and geographic location information of the scenic spot, converting the geographic location information into a matrix form, and generating an initial population of individuals containing multiple itinerary routes through a genetic algorithm.

[0081] By considering factors such as the scenic level, opening hours, ticket prices, previous passenger flow, and accommodation conditions of the scenic spot, a fitness function is established to evaluate the quality of each individual in the initial population. Specifically, the higher the value of the fitness function, the better the individual's travel itinerary plan.

[0082] According to the fitness value, excellent individuals are selected for gene crossover and random mutation to generate mutant populations, which provide a basis for subsequent optimization iterations. The transfer learning technology is used to extract the computational experience of the genetic algorithm from the mutant population and introduce it into the reinforcement learning mechanism, which helps to accelerate the training process of the reinforcement learning model and improve the performance of the model.

[0083] Optimization Iteration Module: Based on transfer learning, the computational experience of the genetic algorithm is extracted from the mutant population, and a reinforcement learning model is established. The model contains a reward mechanism function to represent the expected reward of the state-action, and training optimization is performed through fine-tuning methods.

[0084] Combined with the real-time feedback from travelers, multiple iterative calculations are performed. Each iterative calculation updates the strategy based on the difference and reward of the previous one to optimize the travel route in real time. The formula of iterative calculation expresses the process of state transfer and action selection. Through multiple iterative calculations, the strategy is continuously updated until the convergence condition is reached or the preset number of iterations is reached. Finally, a real-time optimized travel route is obtained.

[0085] Visual feedback module: receives the real-time optimized travel route from the optimization iteration module, and outputs the optimal travel itinerary plan in a visual form, including map display, sequence of attractions, time arrangement and other detailed information, which is convenient for travelers to view and understand.

[0086] This system generates an initial travel itinerary plan through a genetic algorithm and uses a reinforcement learning mechanism for real-time optimization. During the optimization process, the system fully considers various factors of scenic spots, such as scenic grade, opening hours, ticket prices, etc., as well as real-time feedback from travelers. Through multiple iterative calculations, the system continuously updates strategies and finally obtains the optimal travel itinerary plan, which is presented to travelers in a visual form. The travel itinerary planning system based on genetic algorithm and reinforcement learning provided by the present invention can provide travelers with personalized, real-time travel planning services, thereby improving the quality of travel experience.

[0087] The above is based on the ideal embodiment of the present invention. Through the above description, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention. The technical scope of the present invention is not limited to the content in the specification, and the technical scope must be determined according to the scope of the claims.

Claims

1. A travel itinerary planning system based on genetic algorithm, comprising: Big data module, initialization module, optimization iteration module and visual feedback module; characterized in that, The big data module stores basic information of each scenic spot, including the name, geographical location, scenic grade, opening hours, ticket price, previous passenger flow, and accommodation conditions of the scenic spot; The initialization module receives a travel itinerary planning request from a traveler based on a genetic algorithm, extracts scenic spot information that meets the planning request from the big data module, pre-processes and classifies the scenic spot information, generates an initial population, establishes a fitness function in the initialization module, and is used to calculate and evaluate the quality of each individual in the initial population; selects excellent individuals according to the fitness value, crosses some genes of the excellent individuals to generate new individuals, and randomly mutates the new individuals to generate a mutant population; The optimization iteration module migrates the variant population to a new population based on the reinforcement learning of transfer learning, and introduces a reinforcement learning mechanism to optimize the travel route in real time after multiple iterative calculations based on the real-time feedback of travelers; The visualization feedback module receives the real-time optimized travel route obtained from the optimization iteration module, and outputs the optimal travel itinerary planning in a visualized form.

2. A travel itinerary planning system based on genetic algorithm according to claim 1, characterized in that: The process of establishing the initial population includes: obtaining the name and geographical location information of the scenic spot, preprocessing the geographical location information, converting the coordinate information of the geographical location into matrix information using standard Euclidean distance, and generating population individuals containing multiple travel routes.

3. A travel itinerary planning system based on genetic algorithm according to claims 1 and 2, characterized in that: The initialization module establishes a fitness function by considering the scenic grade, opening hours, ticket price, previous passenger flow and accommodation conditions of the scenic spot, and calculates the quality of the individuals in the population. The formula expression of the fitness function is: F(i)=α*S(i)+β*T(i)-γ*P(i)-δ*C(i)+ε*A(i) Where F(i) is the fitness value of the population individual, which is used to evaluate the quality of the population individual; α, β, γ, δ, and ε are the weight coefficients of scenery grade, opening hours, ticket price, previous passenger flow, and accommodation conditions, respectively; S(i) is the scenery grade score of the i-th population individual in the initial population; T(i) is the convenience score of the opening hours of the i-th population individual in the initial population; P(i) is the score of the scenic spot ticket price of the i-th population individual in the initial population; C(i) is the score of the previous passenger flow of the i-th population individual in the initial population; A(i) is the score of the accommodation conditions of the i-th population individual in the initial population.

4. A travel itinerary planning system based on genetic algorithm according to claim 1, characterized in that: The reinforcement learning process based on transfer learning includes: using transfer learning technology to extract the computational experience of the genetic algorithm from the mutant population, and transferring it into the reinforcement learning, establishing a reinforcement learning model based on the computational experience, training and optimizing the reinforcement learning model through a fine-tuning method, and establishing a reward mechanism function in the reinforcement learning model, wherein the reward mechanism function includes a state value function and an action value function, which are used to represent the expected reward of the state-action; combining the real-time feedback of the traveler, performing multiple iterative calculations, and each iterative calculation updates the strategy according to the difference and reward of the previous one, so as to obtain a real-time optimized travel route.

5. A travel itinerary planning system based on genetic algorithm according to claims 1 and 4, characterized in that: The formula calculation expression of the state value function is: The formula calculation expression of the action value function is: Where V Φ (s) represents the expected return obtained by following the strategy Φ starting from state s; Q Φ (s,a) represents the expected return after starting from state s, executing action a, and then following strategy Φ; s t and a t are the state and action at time t, r(s t ,a t ) is in state s t Execute action a t The immediate reward obtained after Φ is the expectation under strategy Φ; η t The discount rate used to calculate the present value of future rewards, between 0 and 1.

6. A travel itinerary planning system based on genetic algorithm according to claims 1, 4 and 5, characterized in that: The formula calculation expression for multiple iterative calculations is: V Φ (s)=E Φ [r(s,a)+ηV Φ (s′)|s] Q Φ (s,a)=E Φ [r(s,a)+ηQ Φ (s′,a′)|s,a] Where s′ is the new state after state s executes action a, that is, the result after each iteration; a′ is the action selected according to strategy Φ in state s′; the meanings of the other variables are the same as above.

7. A travel itinerary planning system based on genetic algorithm according to claims 5 and 6, characterized in that: After multiple iterations, the optimal value formula of the state value function is calculated as follows: Where V * (s) is the state value function with the largest value among all strategies Φ; Q * (s,a) is the action-value function with the largest value among all strategies Φ; the meanings of the remaining variables are the same as above.

8. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the system according to any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the system according to any one of claims 1 to 7 are implemented.

Images