Optimization method, system, device and medium for catastrophe reinsurance contract

Abstract

This invention provides a method, system, device, and medium for optimizing catastrophe reinsurance contracts, belonging to the insurance field. The method includes: constructing a multi-agent system; employing a mathematical optimization model, combined with core clauses in the reinsurance contract, to design a dynamically optimized reinsurance contract mechanism to optimize the contract terms; designing a collaboration and game mechanism among the agents, and using multi-agent reinforcement learning to dynamically optimize the game strategy to adjust the contract terms in real time. A strategy combining game theory and reinforcement learning is used to optimize the contract terms between the insurance company and the reinsurance company. The game mechanism helps the agents find the optimal balance between competition and cooperation, while reinforcement learning enables the agents to continuously adjust their strategies in a dynamic market. This ensures that the reinsurance contract terms can be automatically optimized in response to market changes, disaster events, and changes in regulatory requirements. The agents optimize the contract terms through a real-time feedback mechanism, ensuring a balance between risk sharing and profitability.

Description

Technical Field

[0001] This invention belongs to the field of insurance technology, and specifically relates to an optimization method, system, device and medium for catastrophe reinsurance contracts. Background Technology

[0002] Catastrophe reinsurance is an insurance product that provides risk sharing for insurance companies, primarily addressing the substantial economic losses caused by natural disasters such as earthquakes, typhoons, and floods, or other large-scale unforeseen events. Traditional catastrophe reinsurance contracts are typically negotiated and formulated between insurance and reinsurance companies based on historical data, statistical models, and manual pricing strategies. However, with global climate change and the increasing frequency of disasters, traditional catastrophe reinsurance models face many new challenges and shortcomings.

[0003] Traditional catastrophe reinsurance contracts typically rely on historical data and disaster models for risk assessment. However, this approach depends on historical disaster events, neglecting emerging and yet-to-emerge risk factors. With climate change and continuous changes in the social environment, past risk patterns may no longer accurately predict future disaster risks. Therefore, traditional risk assessment methods are outdated and struggle to effectively address new risk challenges. Currently, catastrophe reinsurance contract terms are usually set in a one-off manner, lacking sufficient flexibility and dynamic adjustment mechanisms. As market conditions, risk exposure, and the frequency of disasters change, contract terms may need to be adjusted, but traditional methods cannot respond quickly to these changes, potentially leading to contract terms that are mismatched with market demands, thus affecting the effectiveness and fairness of the contract. Furthermore, they ignore the cooperation and game theory among multiple participants, resulting in inefficient decision-making processes. Summary of the Invention

[0004] The purpose of this invention is to provide an optimization method, system, device, and medium for catastrophe reinsurance contracts, which can solve, in whole or in part, the technical problems existing in the prior art, such as the lag in evaluation methods, the lack of dynamic adjustment mechanisms, and the neglect of cooperation and game among multiple participants.

[0005] In a first aspect, embodiments of this application provide a method for optimizing a catastrophe reinsurance contract, comprising:

[0006] Constructing multi-agent systems;

[0007] A mathematical optimization model is adopted, and combined with the core clauses in the reinsurance contract, a dynamic optimization reinsurance contract mechanism is designed to optimize the reinsurance contract clauses. The optimization of the reinsurance contract mechanism includes a reinsurance contract parameter optimization strategy and multi-agent collaborative optimization of contract parameters.

[0008] The design incorporates collaboration and game-playing mechanisms among the agents, and employs multi-agent reinforcement learning to dynamically optimize game strategies and adjust contract terms in real time. These mechanisms include the game-playing mechanism between the insurance company and the reinsurance company, as well as the agent collaboration mechanism.

[0009] Optionally, the constructed multi-agent system includes at least an insurance company agent, a reinsurance company agent, a risk assessment agent, a regulatory compliance agent, and an external environment agent.

[0010] Optional reinsurance contract parameter optimization strategies include: reinsurance ratio optimization, reinsurance rate optimization, and deductible optimization;

[0011] The reinsurance ratio is optimized according to the following formula:

[0012]

[0013] In the formula, For the optimal reinsurance ratio, This indicates penalties for failing to meet regulatory requirements regarding capital adequacy ratios. The penalty coefficient is used to balance profit and the risk of SCR violation. The expected profit of the product line.

[0014] Optimize reinsurance rates using the following formula:

[0015]

[0016] In the formula, P* is the optimal reinsurance rate, E[L|L>d] is the expected loss after exceeding the deductible d, and ζ is the profit loading factor of the reinsurance company.

[0017] Optimize the deductible using the following formula:

[0018]

[0019] In the formula, To be the optimal deductible, It is the degree of risk exposure. It is a risk adjustment parameter. The expected profit of the product line.

[0020] Optional, multi-agent cooperative optimization contract parameters include:

[0021] For insurance company agents, the reinsurance ratio is adjusted based on capital adequacy ratio and market conditions;

[0022] For reinsurance companies' intelligent agents, reinsurance rates and deductibles are adjusted according to market competition.

[0023] For regulatory compliance agents, provide real-time compliance checks to ensure that capital adequacy ratios meet regulatory requirements;

[0024] For risk assessment agents, market risk forecasts are provided, which affect decisions regarding reinsurance ratios and deductibles;

[0025] For intelligent agents targeting the external environment, monitoring market economic conditions influences rate adjustments.

[0026] Optional game mechanisms between insurance companies and reinsurance companies include:

[0027] Under stable market conditions, insurance companies and reinsurance companies optimize contracts through Stackelberg reinforcement learning. The reinsurance company first formulates reinsurance contract parameters with the goal of maximizing profits. These reinsurance contract parameters include reinsurance rates, reinsurance ratios, and deductibles.

[0028] After the reinsurer sets the contract parameters, the insurance company selects the optimal reinsurance purchase strategy based on its own capital adequacy ratio and market risk.

[0029] In the event of severe market volatility or catastrophic events, insurance and reinsurance companies choose to cooperate and share profits according to the Shapley value distribution mechanism.

[0030] Optional multi-agent reinforcement learning methods include:

[0031] When making decisions, each agent defines a state vector based on the market environment, contract terms, and historical data.

[0032] For insurance company agents, set reinsurance ratios and whether to accept reinsurance contracts; for reinsurance company agents, set reinsurance rates and deductibles; for risk assessment agents, determine whether to renegotiate reinsurance contract terms; for regulatory compliance agents, determine whether to change the minimum solvency adequacy ratio; for external environment agents, determine whether to change risk exposure.

[0033] Design an agent reward function and adopt an Actor-Critic architecture to optimize contract parameters. The Actor network is used to generate the optimal contract policy, and the Critic network is used to evaluate the performance of the current policy in long-term returns and adjust the policy accordingly.

[0034] Initialize the neural network for each agent, the ReplayBuffer for storing the agent's historical interaction data, and set the learning rate and discount factor;

[0035] Each agent interacts with the environment based on the current policy, obtains the state, selects actions, receives immediate rewards and the next state, and stores the interaction results in the Replay Buffer;

[0036] Training is performed by randomly sampling from the Replay Buffer, calculating the Q-value update for each agent, updating the agent's policy using the deterministic policy gradient, and estimating the Q-value using the target network. The training process terminates when the agent's policy converges after multiple training epochs and the long-term reward of each agent reaches its optimum.

[0037] Optionally, the strategy can be updated according to the following formula:

[0038]

[0039]

[0040] In the formula, It's the learning rate. It is a state-action value function. It is about gradient, It is about The gradient.

[0041] Secondly, embodiments of this application also provide an optimization system for catastrophe reinsurance contracts, comprising:

[0042] Building blocks are used to construct multi-agent systems;

[0043] The design unit is used to design a dynamically optimized reinsurance contract mechanism by employing a mathematical optimization model and combining the core clauses in the reinsurance contract to optimize the reinsurance contract terms. The optimization of the reinsurance contract mechanism includes a reinsurance contract parameter optimization strategy and multi-agent collaborative optimization of contract parameters.

[0044] The optimization unit is used to design the cooperation and game mechanism between the agents and to dynamically optimize the game strategy using a multi-agent reinforcement learning method to adjust the contract terms in real time. The cooperation and game mechanism between the agents includes the game mechanism between the insurance company and the reinsurance company, as well as the agent coordination mechanism.

[0045] Thirdly, embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the above-described method for optimizing a catastrophe reinsurance contract.

[0046] Fourthly, embodiments of this application also provide a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the optimization method for the catastrophe reinsurance contract described above.

[0047] As can be seen from the above technical solutions, the present invention has the following advantages:

[0048] The method, system, equipment, and medium for optimizing catastrophe reinsurance contracts provided in this application employ a strategy combining game theory and reinforcement learning to optimize the contract terms between insurance companies and reinsurance companies. The game theory mechanism helps the agent find the optimal balance between competition and cooperation, while reinforcement learning enables the agent to continuously adjust its strategy in a dynamic market. This ensures that the reinsurance contract terms can be automatically optimized in response to market changes, disaster events, and changes in regulatory requirements. The agent optimizes the contract terms through a real-time feedback mechanism, ensuring a balance between risk sharing and profitability. Attached Figure Description

[0049] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A flowchart illustrating an optimization method for a catastrophe reinsurance contract provided in an embodiment of the present invention;

[0051] Figure 2 The following is a construction process of a multi-agent reinforcement learning algorithm provided in an embodiment of the present invention;

[0052] Figure 3 A schematic diagram of the structure of an optimization system for a catastrophe reinsurance contract provided in an embodiment of the present invention;

[0053] Figure 4 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0054] Various embodiments of this disclosure will be described more fully in the following detailed description. This disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of this disclosure to the specific embodiments disclosed herein, but rather this disclosure should be understood to cover all adjustments, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of this disclosure.

[0055] In the following, the terms “comprising” or “may include”, which may be used in various embodiments of this disclosure, indicate the presence of the disclosed functions or operations and do not limit the addition of one or more functions or operations. Furthermore, as used in various embodiments of this disclosure, the terms “comprising,” “having,” and their cognates are intended only to indicate a specific feature, number, step, operation, or combination of the foregoing and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, or combinations of the foregoing, or the possibility of adding one or more features, numbers, steps, operations, or combinations of the foregoing.

[0056] In various embodiments of this disclosure, the expression "or" or "at least one of A and / or B" includes any combination or all combinations of the words listed simultaneously. For example, the expression "A or B" or "at least one of A and / or B" may include A, may include B, or may include both A and B.

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] See Figure 1 The diagram shows a flowchart of an optimization method for a catastrophe reinsurance contract in a specific embodiment, including the following execution steps:

[0059] Step 100: Construct a multi-agent system.

[0060] Specifically, the constructed multi-agent system includes at least an insurance company agent, a reinsurance company agent, a risk assessment agent, a regulatory compliance agent, and an external environment agent.

[0061] Specifically, the process of constructing a multi-agent system includes:

[0062] For the insurance company's intelligent agent, the goals and functions are defined as follows: Goal: To manage insurance business, including underwriting and claims settlement, to achieve profitability and maintain stable operations. Functions: Receive customer insurance applications, assess risks to determine underwriting and premiums; process claims, calculate and pay compensation; manage insurance product inventory and formulate insurance product strategies. Knowledge and Data Storage: Insurance Product Information Database: Stores detailed terms, insured amount ranges, rate tables, etc., for various insurance products. Customer Information Database: Records basic customer information, insurance history, claims records, etc. Risk Assessment Model: Used to assess customer risk levels and determine premium levels. Decision-Making Mechanism Construction: Insurance Application Decision: Based on customer risk assessment results, product inventory, and market strategies, decides whether to accept the insurance application; if accepted, determines the premium. Claims Settlement Decision: Calculates and executes compensation based on insurance contract terms, claims application materials, and investigation results. Communication Interface Design: Customer Interaction Interface: Receives insurance applications and claims requests, and provides feedback on processing results. Reinsurance Company Communication Interface: Negotiates reinsurance business, such as reinsurance ratios and premium payments. Interface with the risk assessment agent: Obtain customer risk assessment reports. Interface with the regulatory compliance agent: Receive regulatory update notifications to ensure business compliance.

[0063] For the reinsurance company's intelligent agent, the objectives and functions are defined as follows: Objective: To obtain revenue and control its own risk exposure by assuming a portion of the insurance company's risk. Function: To assess the reinsurance needs of insurance companies, determine reinsurance terms and rates; manage the reinsurance portfolio, diversify and optimize risk; and provide reinsurance claims support. Knowledge and Data Storage: Reinsurance Product Library: Contains the terms, reinsurance ratio range, rate structure, etc. of various reinsurance products. Insurance Company Information Library: Records the business scale, risk status, and historical reinsurance records of cooperating insurance companies. Risk Assessment Model: Used to assess the risks transferred by insurance companies. Decision-Making Mechanism Construction: Reinsurance Decision: Based on the insurance company's risk assessment results, its own risk tolerance, and market conditions, decides whether to accept the reinsurance request and determines the reinsurance terms and rates. Claims Decision: In the event of a reinsurance claim, calculates and pays the compensation amount based on the reinsurance contract terms and the insurance company's claims history. Communication Interface Design: Communication Interface with Insurance Companies: Receives reinsurance applications, provides feedback on reinsurance decision results, and processes claims notifications and settlements. Interface with External Environment Intelligent Agent: Obtains macroeconomic data, industry risk indices, etc., to assist in decision-making.

[0064] For the risk assessment agent, the goals and functions are defined as follows: Goal: To provide accurate risk assessment services to insurance and reinsurance companies. Function: To collect and analyze various risk-related data, and use risk assessment models to assess customer risk, insurance business risk, and reinsurance business risk. Knowledge and Data Storage: Risk Data Warehouse: Collects data from multiple channels, including customer information, industry data, historical claims data, market data, etc. Risk Assessment Model Library: Stores various risk assessment models, such as credit risk assessment models, property risk assessment models, and health risk assessment models. Decision Mechanism Construction: Risk Assessment Process: Based on the input data, selects an appropriate risk assessment model for calculation, generates a risk assessment report, and the report content includes risk level, risk probability, potential loss, etc. Communication Interface Design: Interface with Insurance Companies: Receives customer insurance information and insurance business data, and provides risk assessment reports. Interface with Reinsurance Companies: Receives reinsurance business-related data and provides risk assessment services. Interface with External Environment Agent: Obtains external data that may affect risk assessment, such as natural disaster data and economic data.

[0065] For the regulatory compliance intelligent agent, the objectives and functions are defined as follows: Objective: To ensure that the business activities of all intelligent agents in the insurance industry comply with legal and regulatory requirements. Function: To track and interpret changes in relevant laws and regulations in the insurance industry; to review the compliance of business processes and product terms of insurance companies and reinsurance companies; and to provide compliance training and consulting services. Knowledge and Data Storage: Regulatory Policy Database: Stores domestic and international laws, regulations, and regulatory policy documents related to the insurance industry. Compliance Case Library: Collects and organizes compliance cases in the insurance industry for reference and learning. Decision-Making Mechanism Construction: Compliance Review Process: Based on input business data (such as insurance product terms and business operation processes), reviews the data against the regulatory policy database to determine compliance. If non-compliance issues are found, provides rectification suggestions. Communication Interface Design: Interface with Insurance Companies: Receives information such as insurance product design and business operation processes, and provides feedback on compliance review results. Interface with Reinsurance Companies: Reviewes the compliance of reinsurance business. Interface with External Environment Intelligent Agents: Obtains updated regulatory policy information and updates the regulatory policy database in a timely manner.

[0066] For the external environment intelligent agent, the goals and functions are defined as follows: Goal: To simulate and provide information on the impact of external environmental factors on the insurance industry. Function: To collect and integrate macroeconomic data, industry dynamics, natural disaster data, public opinion, etc.; analyze the impact of this data on the insurance industry and provide relevant information to other intelligent agents. Knowledge and Data Storage: External Data Warehouse: Stores various external data from government departments, industry associations, news media, etc. Environmental Impact Analysis Model: Used to analyze the impact of external data on the insurance industry, such as models of the impact of economic recession on insurance demand, and models of the impact of natural disasters on claims costs. Decision-Making Mechanism Construction: Data Collection and Analysis Process: Regularly collect external data, analyze it using environmental impact analysis models, and generate environmental impact reports. Report content includes economic trend forecasts, industry risk warnings, and natural disaster risk assessments. Communication Interface Design: Interface with Insurance Companies: Provide information on macroeconomics and industry dynamics to help them formulate business strategies. Interface with Reinsurance Companies: Provide information on external environments that may affect reinsurance business. Interface with Risk Assessment Intelligent Agent: Provide external data to assist in risk assessment. Interface with Regulatory Compliance Intelligent Agent: Timely deliver updated information on regulations and policies.

[0067] It is important to note that clearly defining the roles of each agent in a multi-agent system and their optimization objectives lays the foundation for subsequent contract optimization, game theory mechanism design, dynamic adjustment, and real-time feedback mechanisms. Each agent has different objectives, constraints, and interaction methods. Through the combination of multi-agent reinforcement learning (MARL) and game theory mechanisms, they work collaboratively to achieve the final contract optimization and risk control.

[0068] For example, the objectives of the insurance company's intelligent agent are defined as optimizing the insurance company's capital adequacy ratio, reducing claims risk, reducing reinsurance costs, and improving capital utilization efficiency; the decision variable is the reinsurance ratio (…). The decision variable is the reinsurance premium rate. The agent chooses whether or not to accept the reinsurance contract (g). The constraints are meeting regulatory requirements, ensuring a reasonable capital adequacy ratio, and effectively controlling risk exposure. The objective of the reinsurance agent is defined as maximizing premium income and ensuring profit while assuming risk and controlling risk exposure. ), deductible ( The constraints are: ensuring sufficient capital and profitability while providing reasonable reinsurance terms to attract insurance companies. The goal of the risk assessment agent is to provide accurate risk prediction and assessment, assisting insurance and reinsurance companies in adjusting contract terms to ensure appropriate risk sharing; the decision variable is whether to renegotiate reinsurance contract terms; the constraint is to ensure the accuracy and stability of risk assessment based on historical data and market trends. The goal of the regulatory compliance agent is to ensure all contract terms comply with regulatory requirements, avoiding compliance issues for insurance and reinsurance companies and preventing default risks; the decision variable is whether to change the minimum solvency ratio; the constraint is to ensure contract terms meet minimum capital adequacy requirements and monitor the compliance of all operations within the industry. The goal of the external environment agent is to simulate and predict the impact of external environments such as market changes, economic risks, and natural disasters on contract terms and agent decisions; the decision variable is whether to change risk exposure; the constraint is to influence the capital allocation and risk management decisions of insurance and reinsurance companies, but not to directly interfere with contract terms.

[0069] Step 101: Using a mathematical optimization model, and combining the core clauses of the reinsurance contract, design a dynamic optimization reinsurance contract mechanism to optimize the reinsurance contract terms.

[0070] The optimization of the reinsurance contract mechanism includes a reinsurance contract parameter optimization strategy and a multi-agent collaborative optimization of contract parameters.

[0071] The core objective of this step is to optimize the reinsurance contract structure, ensuring that insurance companies can obtain reinsurance coverage at the optimal cost while meeting capital adequacy ratio (SCR) requirements, and simultaneously enabling reinsurance companies to obtain reasonable returns while bearing risks. To this end, a mathematical optimization model is employed, combined with core clauses in the reinsurance contract (reinsurance ratio, deductible, reinsurance rate, etc.), to design a dynamically optimized reinsurance contract mechanism to adapt to market changes and improve the stability and adaptability of the contract.

[0072] For example, reinsurance contract parameter optimization strategies include: reinsurance ratio optimization, reinsurance rate optimization, and deductible optimization.

[0073] Among these factors, insurance companies aim to reduce their own risk exposure while increasing their reinsurance ratio. This may lead to increased reinsurance premiums. The goal is to optimize the reinsurance ratio while ensuring SCR compliance, specifically by optimizing the reinsurance ratio according to the following formula:

[0074]

[0075] In the formula, For the optimal reinsurance ratio, This indicates penalties for failing to meet regulatory requirements regarding capital adequacy ratios. The penalty coefficient is used to balance profit and the risk of SCR violation. The expected profit of the product line.

[0076] If market volatility is high (risk assessment agent issues warning), then Increase, reduce the risk exposure of insurance companies. If the market is stable (claims rate decreases), then This reduces costs and increases insurance company profits.

[0077] Reinsurance companies need to set an optimal premium rate *p* to ensure profitability while avoiding excessively high rates that could reduce market competitiveness. The goal is to achieve the optimal balance between market competition and risk acceptability: optimize the reinsurance premium rate using the following formula:

[0078]

[0079] In the formula, P* is the optimal reinsurance rate, E[L|L>d] is the expected loss after exceeding the deductible d, and ζ is the profit loading factor of the reinsurance company.

[0080] If market demand increases (insurance companies increase their reinsurance needs), then If the market competition intensifies (multiple reinsurance companies vying for market share), then... By lowering rates, reinsurance companies can enhance their competitiveness.

[0081] The deductible d affects the risk-sharing between the insurer and reinsurer, and it is necessary to ensure that the benefits for both are optimal. The goal is to optimize d to balance the interests of the reinsurer and the insurer: optimize the deductible according to the following formula:

[0082]

[0083] In the formula, To be the optimal deductible, It is the degree of risk exposure. It is a risk adjustment parameter. The expected profit of the product line.

[0084] If the risk assessment agent predicts increased market volatility, it increases d, reducing the reinsurer's liability for claims. If the market is stable, it decreases d, increasing the claims undertaken by the insurer and improving its profits.

[0085] For example, multi-agent collaborative optimization of contract parameters includes: for the insurance company agent, adjusting the reinsurance ratio based on the capital adequacy ratio (SCR) and market conditions. For reinsurance agents, adjust reinsurance rates P and deductibles d based on market competition; for regulatory compliance agents, provide real-time compliance checks to ensure capital adequacy ratios (SCR) meet regulatory requirements; for risk assessment agents, provide market risk predictions that affect reinsurance ratios. The decision-making process regarding the deductible d; for intelligent agents in the external environment, monitoring market economic conditions, and influencing rate adjustments.

[0086] Step 102: Design a collaboration and game mechanism among the agents, and use a multi-agent reinforcement learning method to dynamically optimize the game strategy to adjust the contract terms in real time.

[0087] Among them, the cooperation and game mechanism among the intelligent agents includes the game mechanism between insurance companies and reinsurance companies, as well as the collaborative mechanism among intelligent agents.

[0088] To resolve the conflicts of interest and information asymmetry between insurance companies and reinsurance companies, in the reinsurance market, insurance companies need to reduce underwriting risk through reinsurance, while reinsurance companies aim to maximize their profits while controlling risk exposure. Therefore, the contract negotiation process can be modeled as a multi-agent game problem involving both cooperation and competition mechanisms.

[0089] Specifically, the game mechanism between insurance companies and reinsurance companies includes:

[0090] Under stable market conditions, insurance companies and reinsurance companies optimize contracts using Stackelberg reinforcement learning. The reinsurance company first formulates reinsurance contract parameters, including the reinsurance premium rate p and the surrogacy ratio. And the deductible d, its goal is to maximize profit:

[0091]

[0092] In the formula, For the benefit of reinsurance companies. Premiums received by the reinsurance company This is the expected payout amount for the reinsurer.

[0093] After the reinsurer sets the contract parameters, the insurance company selects the optimal reinsurance purchase strategy based on its own capital adequacy ratio and market risk.

[0094]

[0095] In the formula, .

[0096] Insurance companies also use reinforcement learning for optimization, learning the optimal purchasing strategy through deep deterministic policy gradients.

[0097] When market volatility is high or catastrophic events occur, insurance companies and reinsurance companies choose to cooperate and optimize contract terms through collaborative game theory, so that both parties can still profit when the market is unstable and ensure market stability.

[0098] When an insurance company collaborates with a reinsurance company, its total revenue is:

[0099]

[0100] Insurance companies and reinsurance companies share profits according to the Shapley value distribution mechanism:

[0101]

[0102] in, For intelligent agents Profit distribution in cooperation For the collection of intelligent agents participating in the cooperation Total revenue.

[0103] In some implementations, a collaborative mechanism among other intelligent agents includes a risk assessment agent that provides accurate risk assessments for insurance and reinsurance companies, particularly offering strategy adjustment recommendations in the event of extreme risks and catastrophic events. This agent assesses the risk of potential extreme events (such as catastrophic events, financial crises, etc.) based on historical data, market trends, and disaster models, with a reward function ( )for:

[0104]

[0105] in, The probability of extreme events. The expected losses caused by extreme events.

[0106] Regulatory bodies influence the design of contract terms for insurers and reinsurers through minimum capital adequacy ratio (SCR) requirements, and reward functions. )for:

[0107]

[0108] in, The minimum solvency adequacy ratio, For the current capital adequacy ratio, It is a market adjustment factor.

[0109] The external environment intelligent agent simulates factors such as market fluctuations, economic indicators, and natural disasters, affecting the risks faced by insurance and reinsurance companies during contract optimization. Reward function ( )for:

[0110]

[0111] in, For the debt risk of insurance companies, This is related to the debt risk of reinsurance companies.

[0112] In one specific implementation, multiple agents optimize strategies through a game theory mechanism. To enable effective strategy adjustments in complex and dynamic environments, a multi-agent reinforcement learning (MARL) method is employed. MARL combines the self-learning capabilities of reinforcement learning with the theoretical framework of game theory, enabling dynamic strategy optimization among multiple agents and real-time adjustment of contract terms based on market and risk changes to ensure the maximization of interests for all parties. (See [reference needed]). Figure 2 As shown, multi-agent reinforcement learning includes the following steps:

[0113] S200: When making decisions, each agent defines a state vector based on the market environment, contract terms, and historical data.

[0114] For example, the state vector is defined as follows:

[0115]

[0116] in, This represents the current capital adequacy ratio of insurance companies. For the matching of assets and liabilities, For liquidity coverage ratio, As a market volatility indicator, Historical claims rate Due to current regulatory policy constraints, Claim prediction values ​​provided to the risk assessment agent.

[0117] S201: For insurance company agents, set the reinsurance ratio and whether to accept reinsurance contracts; for reinsurance company agents, set the reinsurance rate and deductible; for risk assessment agents, whether to renegotiate reinsurance contract terms; for regulatory compliance agents, whether to change the minimum solvency adequacy ratio; for external environment agents, whether to change risk exposure.

[0118] S202: Design the agent reward function and optimize the contract parameters.

[0119] Specifically, the global reward function is used to optimize contract parameters while taking into account the objectives of insurance companies, reinsurance companies, and regulatory agencies:

[0120]

[0121] in, These are the weighting coefficients.

[0122] S203: Adopts an Actor-Critic architecture, in which the Actor network is used to generate the optimal contract strategy, the Critic network is used to evaluate the performance of the current strategy in long-term returns, and the strategy is adjusted.

[0123] Specifically, the Actor network operates as follows when generating contract strategies: Market information, supplier information, and platform information are input into the input layer. After the input information enters the hidden layer, feature extraction is performed. Each neuron processes a portion of the input information, extracting higher-level features. For example, for market price fluctuations and historical supplier price data, neurons can extract price trend features to determine whether prices are rising, falling, or stable. Nonlinear transformations are applied to the extracted features using nonlinear activation functions (such as the ReLU function). For instance, when market demand is clearly rising and platform inventory is low, the network can more accurately capture the information that a proactive procurement strategy is needed in this situation after the nonlinear transformation. Features extracted by different neurons are fused in the hidden layer to form a more comprehensive and representative feature vector. These fused features are passed between hidden layers for further processing and analysis, providing a richer information foundation for the output layer to generate contract strategies. After processing by the hidden layer, the output layer calculates the probability distribution of different contract strategies (actions) to be taken in the current state. Assume that the contract strategy includes several dimensions such as price strategy, delivery time strategy, and quality standard strategy. For pricing strategies, the output might be the probability of choosing a low-price, medium-price, or high-price strategy under the current circumstances. For delivery time strategies, the output might be the probability of choosing a short, medium, or long delivery period. Based on the output probability distribution, a sampling method (such as roulette wheel selection) is used to determine the final contract strategy. For example, if the probability of choosing a low-price strategy is 0.6, the probability of choosing a medium-price strategy is 0.3, and the probability of choosing a high-price strategy is 0.1, then in multiple samplings, the low-price strategy will be chosen approximately 60% of the time. Simultaneously, by combining the strategy probabilities from other dimensions such as delivery time and quality standards, a complete contract strategy is ultimately formed.

[0124] Similarly, the Critic network receives state information about the current environment. This includes market supply and demand, such as the range of market price fluctuations and predicted market demand; and the contract strategy chosen by the Actor network, i.e., action information. This includes the purchase price range, delivery time window, stringency of quality acceptance standards, and specific after-sales service requirements specified in the contract. This action information is also converted into appropriate numerical representations and input into the network. The input state and action information first enters the hidden layers of the Critic network. In the hidden layers, the network performs feature extraction on the input information. Each neuron performs a weighted summation of a portion of the input information and processes it through a non-linear activation function (such as ReLU) to extract more meaningful features. As information is continuously passed and processed in the hidden layers, the features extracted by different neurons gradually merge. Through the computation of multiple hidden layers, the network integrates various features to form a comprehensive understanding of the current state-action pair. Finally, the output layer of the network outputs a single numerical value, which represents the expected long-term value (also known as value estimation) of taking this contract strategy in the current state. For example, if current market demand is strong, and the contract strategy offers reasonable pricing and short delivery times, the Critic network might output a high value estimate, indicating that the strategy is likely to yield high returns in the long run. After contract execution, the platform receives an actual reward based on the actual returns. The Critic network updates its parameters using optimization algorithms (such as gradient descent) based on the difference between the predicted value and the actual reward. The goal is to make the network's predicted value more accurately reflect actual long-term returns. For example, if the predicted value consistently exceeds the actual reward, the network adjusts the connection weights between neurons to lower future predicted values; conversely, it increases predicted values. The Critic network's evaluation results are fed back to the Actor network. The Actor network adjusts its contract strategy generation based on the Critic network's feedback. If the Critic network deems a strategy valuable, the Actor network increases the probability of adopting that strategy in similar situations; if the value is low, it decreases the probability, thus gradually learning contract strategies that yield higher long-term returns.

[0125] For example, a multi-agent reinforcement learning method based on an Actor-Critic architecture is employed to improve the game strategy of insurance companies and reinsurance companies in contract negotiations: The Actor network generates the optimal contract strategy (e.g., reinsurance ratio, reinsurance rate). The Critic network evaluates the performance of the current strategy in long-term returns and adjusts the strategy accordingly.

[0126] S204: Initialize the neural network for each agent, the ReplayBuffer for storing the agent's historical interaction data, and set the learning rate and discount factor.

[0127] S205: Each agent interacts with the environment based on the current policy, obtains the state, selects an action, receives an immediate reward and the next state, and stores the interaction results in the Replay Buffer.

[0128] It should be understood that the interaction result includes state, action, reward, and the next state.

[0129] S206: Randomly sample from the Replay Buffer for training, calculate the Q-value update for each agent, update the agent's policy using the deterministic policy gradient, and use the target network for Q-value estimation; the training process terminates when the agent's policy converges after multiple training epochs and the long-term reward of each agent reaches its optimum.

[0130] Specifically, the strategy is updated according to the following formula:

[0131]

[0132]

[0133] In the formula, It's the learning rate. It is a state-action value function. It is about gradient, It is about The gradient.

[0134] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0135] In some implementations, dynamic adjustment and real-time feedback mechanisms are designed to ensure that reinsurance contracts can adaptively optimize in response to changes in the market, risk, and regulations. This embodiment emphasizes real-time adjustment and dynamic feedback of contract parameters, and utilizes a multi-agent collaborative mechanism combined with the influence of reinforcement learning on contract optimization to achieve a long-term stable optimal contract strategy.

[0136] 1. Strengthen the feedback mechanism driven by learning:

[0137] During contract adjustment, this embodiment employs a reinforcement learning-driven dynamic feedback mechanism to ensure that contract terms are optimized in real time under different market environments. The agent's decisions are not only based on the current market environment but also optimized and adjusted according to real-time feedback to ensure that the contract is always in an optimal state.

[0138] (1) Fine-tuning of reinforcement learning:

[0139] The core of reinforcement learning is learning the optimal strategy. However, due to the constantly changing market environment, agents need to fine-tune their strategies to adapt to new market situations. Each adjusted contract term enters the feedback loop, optimizing decisions in the following ways:

[0140] Fast convergence: Reinforcement learning avoids over-adjustment by fine-tuning the strategy, ensuring that the optimal contract terms are reached quickly.

[0141] Exploration vs. Utilization Balance: When choosing an adjustment strategy, the agent selects either exploration (trying new strategies) or utilization (optimizing based on existing strategies) based on current market feedback.

[0142] (2) Real-time feedback on contract adjustments:

[0143] After each contract adjustment, the AI ​​will adjust its strategy in real time based on actual market feedback (such as changes in insurance company profits, reinsurance company payout risks, and capital adequacy ratios). This process includes:

[0144] Short-term feedback: Monitor market performance immediately after contract adjustments and adjust reinsurance ratios, rates, or deductibles accordingly.

[0145] Long-term feedback: In the long-term market environment, monitor the effectiveness of contract strategies and optimize the adjustment process through reinforcement learning.

[0146] 2. Reinforce learning and agent collaboration:

[0147] The reinforcement learning in this embodiment is not limited to the optimization of a single agent, but emphasizes the collaborative optimization of multiple agents. The actions of each agent will affect the decisions of other agents, thereby achieving the optimal strategy for the entire system.

[0148] (1) Collaborative decision-making mechanism:

[0149] The risk assessment agent influences contract parameters by calculating future risks and making adjustment suggestions to insurance and reinsurance companies.

[0150] The regulatory compliance agent monitors the capital adequacy ratio (SCR) and reminds insurance and reinsurance companies to adjust their contract strategies to ensure compliance with regulatory requirements.

[0151] The external environment agent adjusts its market behavior based on factors such as macroeconomic changes and market risks, and provides market adaptation suggestions to other agents through feedback mechanisms.

[0152] (2) Real-time collaborative adjustment:

[0153] Reinforcement learning optimizes the strategies of each agent through multi-agent collaboration. In each round of contract adjustment, the behavior of each agent (such as the magnitude and timing of contract adjustments) is collaboratively optimized to ensure market stability and the long-term optimality of the contract.

[0154] 3. Stability and compliance of contract adjustments:

[0155] Stability is crucial during contract optimization, especially when market volatility is high. The dynamic adjustment mechanism ensures the stability and compliance of contract adjustments through the following methods:

[0156] Smooth Transition: Contract parameter adjustments are gradual to avoid drastic fluctuations. A reinforcement learning-based smooth update mechanism ensures a stable transition of adjusted contract terms, gradually adapting to the new market environment.

[0157] Compliance assurance: Through real-time monitoring by the regulatory compliance intelligence agent, we ensure that contracts always comply with the latest regulatory requirements (such as capital adequacy ratio, solvency, etc.) during the adjustment process.

[0158] Compared to existing technologies, it has the following significant advantages:

[0159] 1. Achieve dynamic optimization of contract terms:

[0160] By employing multi-agent reinforcement learning, reinsurance contract terms (such as compensation amounts, deductibles, and risk-sharing ratios) can be adjusted in real time based on dynamic information such as market changes, disaster risk assessments, and historical claims data. This ensures that reinsurance companies and insurance companies can always make optimal decisions in a constantly changing risk environment. The adaptive nature of reinforcement learning automates the contract term optimization process without human intervention, thereby improving efficiency and reducing human bias.

[0161] 2. Multi-agent cooperation and game theory optimization:

[0162] By designing a multi-agent system, different roles (such as reinsurance companies, insurance companies, and disaster risk assessment agents) can collaborate and engage in game theory, thereby achieving an optimal balance among the interests of all parties. This collaborative mechanism can effectively reduce the risk of decision-making errors by a single agent while improving the overall efficiency of the system. Introducing a game theory mechanism among agents allows each agent to find the best strategy between cooperation and competition, effectively improving the stability and robustness of the system.

[0163] 3. Improve risk management capabilities:

[0164] By monitoring and assessing factors such as disaster risks and market changes in real time, risk exposure can be dynamically adjusted, optimizing risk allocation in reinsurance contracts. This enables reinsurance companies to better manage and diversify risk, thereby reducing potential losses. Unlike traditional static contract design, this invention allows for timely adjustments to contract terms before disaster events occur and continuous optimization based on real-time feedback, improving the flexibility and responsiveness of risk management.

[0165] 4. Improve decision-making efficiency and automation:

[0166] Through the self-learning and optimization of reinforcement learning agents, decisions can be quickly adjusted according to environmental changes, significantly improving the efficiency of reinsurance contract term optimization. The system's automation level is greatly enhanced, reducing the need for human intervention, thereby lowering labor costs and avoiding human decision-making errors.

[0167] 5. Adapt to the ever-changing market environment:

[0168] Multi-agent reinforcement learning systems possess strong adaptability, enabling them to cope with complex market environments and volatile disaster risks. When market conditions change or disaster events occur, the system can promptly adjust its decisions based on new information, ensuring that contract terms always adapt to changing circumstances. It can adaptively adjust to different types of disaster events (such as storms and earthquakes) and optimize risk-sharing strategies, thereby improving the accuracy and applicability of insurance products.

[0169] 6. Scalability and flexibility:

[0170] Different agent roles can be expanded according to actual needs to adapt to the optimization requirements of reinsurance contracts of different sizes and types. For example, roles such as disaster risk assessment agents and different types of insurance companies can be added as needed to ensure that the system achieves maximum efficiency in multi-party collaboration. The framework is flexible and can be adjusted and optimized for different insurance products and reinsurance needs, making it applicable to a wide range of financial risk management fields.

[0171] 7. Enhanced transparency and explainability:

[0172] By combining reinforcement learning and game theory, the decision-making process is highly transparent, clearly tracing the basis and motivations behind each decision. It not only outputs optimal contract terms but also provides the decision-making rationale and risk assessment process, helping relevant decision-makers understand and accept the system's optimized results.

[0173] like Figure 3As shown, the following are embodiments of the catastrophe reinsurance contract optimization system provided in this disclosure. It belongs to the same inventive concept as the catastrophe reinsurance contract optimization method in the above embodiments. For details not described in detail in the embodiments of the catastrophe reinsurance contract optimization system, please refer to the embodiments of the above catastrophe reinsurance contract optimization method.

[0174] An optimization system for catastrophe reinsurance contracts includes:

[0175] Building blocks are used to construct multi-agent systems;

[0176] The design unit is used to design a dynamically optimized reinsurance contract mechanism by employing a mathematical optimization model and combining the core clauses in the reinsurance contract to optimize the reinsurance contract terms. The optimization of the reinsurance contract mechanism includes a reinsurance contract parameter optimization strategy and multi-agent collaborative optimization of contract parameters.

[0177] The optimization unit is used to design the cooperation and game mechanism between the agents and to dynamically optimize the game strategy using a multi-agent reinforcement learning method to adjust the contract terms in real time. The cooperation and game mechanism between the agents includes the game mechanism between the insurance company and the reinsurance company, as well as the agent coordination mechanism.

[0178] Figure 4 This is a schematic diagram of the hardware structure of an electronic device that implements various embodiments of the present invention.

[0179] The method for optimizing catastrophe reinsurance contracts provided in this application can be applied to electronic devices. Those skilled in the art will understand that the electronic device structure involved in the embodiments of this invention does not constitute a limitation on the electronic device. An electronic device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements. In the embodiments of this invention, the electronic device includes, but is not limited to, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of this application described and / or claimed herein.

[0180] Electronic devices may include processors, external memory interfaces, internal memory, universal serial bus (USB) interfaces, charging management modules, power management modules, batteries, wireless communication modules, audio modules, speakers, microphones, sensor modules, buttons, cameras, displays, and SIM card interfaces, etc.

[0181] It is understood that the structures illustrated in the embodiments of this application do not constitute a specific limitation on the electronic device. In other embodiments of this application, the electronic device may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0182] A processor may include one or more processing units, such as: a central processing unit (CPU), an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural network processing unit (NPU). Different processing units may be independent devices or integrated into one or more processors.

[0183] The processor can serve as the nerve center and command center of an electronic device. The controller can generate operation control signals based on the instruction opcode and timing signals to control the fetching and execution of instructions.

[0184] The processor may also include memory for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. This memory can store instructions or data that the processor has just used or that are used repeatedly. If the processor needs to use the instruction or data again, it can retrieve it directly from this memory. This avoids repeated accesses, reduces processor latency, and thus improves system efficiency.

[0185] An external storage interface (ESI) can be used to connect external memory cards, such as microSD cards, to expand the storage capacity of electronic devices. The external memory card communicates with the processor through the ESI to perform data storage functions, such as saving music and video files on the external memory card.

[0186] Internal memory can be used to store computer executable program code, which includes instructions. The processor executes various functional applications and data processing of electronic devices by running the instructions stored in internal memory. Internal memory can include a program storage area and a data storage area. Internal memory can include high-speed random access memory, and can also include non-volatile memory, such as at least one disk storage device, flash memory device, universal flash storage (UFS), etc.

[0187] Wireless communication functionality in electronic devices can be achieved through antennas, wireless communication modules, modem processors, and baseband processors.

[0188] Wireless communication modules can provide solutions for wireless communication applications in electronic devices, including wireless local area networks (WLANs) (such as wireless fidelity (Wi-Fi) networks), Bluetooth (BT), global navigation satellite system (GNSS), frequency modulation (FM), near field communication (NFC), and infrared (IR) technologies.

[0189] Electronic devices can implement audio functions through audio modules, speakers, receivers, microphones, headphone jacks, and application processors.

[0190] Electronic devices can achieve shooting functions through ISPs, cameras, video codecs, GPUs, displays, and application processors.

[0191] Electronic devices can achieve display functions through GPUs, displays, and application processors.

[0192] A GPU is a microprocessor for image processing, connected to the display screen and application processor. GPUs are used to perform mathematical and geometric calculations for graphics rendering. A processor may include one or more GPUs, which execute program instructions to generate or modify display information.

[0193] A display screen is used to display images, videos, etc. A display screen includes a display panel.

[0194] The storage medium provided in this application stores a program product that enables an optimized method for implementing catastrophe reinsurance contracts.

[0195] The optimization method for catastrophe reinsurance contracts includes: constructing a multi-agent system; employing a mathematical optimization model, combined with the core clauses of the reinsurance contract, to design a dynamically optimized reinsurance contract mechanism to optimize the reinsurance contract terms, wherein the optimization mechanism includes a reinsurance contract parameter optimization strategy and multi-agent collaborative optimization of contract parameters; designing a cooperation and game mechanism among the agents, and employing a multi-agent reinforcement learning method to dynamically optimize the game strategy to adjust the contract terms in real time, wherein the cooperation and game mechanism among the agents includes a game mechanism between the insurance company and the reinsurance company, as well as a cooperation mechanism among the agents.

[0196] In some possible implementations, the subject matter of this disclosure, namely, the method and system for optimizing catastrophe reinsurance contracts, can be implemented as a program product comprising program code that, when run on a terminal device, causes the terminal device to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure.

[0197] The storage medium disclosed herein may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0198] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An optimization method for catastrophe reinsurance contracts, characterized in that, include: Constructing multi-agent systems; A mathematical optimization model is adopted, and combined with the core clauses in the reinsurance contract, a dynamic optimization reinsurance contract mechanism is designed to optimize the reinsurance contract clauses. The optimization of the reinsurance contract mechanism includes a reinsurance contract parameter optimization strategy and multi-agent collaborative optimization of contract parameters. Design a collaboration and game mechanism among the agents, and use a multi-agent reinforcement learning method to dynamically optimize the game strategy to adjust the contract terms in real time. The collaboration and game mechanism among the agents includes the game mechanism between the insurance company and the reinsurance company, as well as the agent coordination mechanism. The constructed multi-agent system includes at least an insurance company agent, a reinsurance company agent, a risk assessment agent, a regulatory compliance agent, and an external environment agent. The construction process of multi-agent systems includes: For different intelligent agents, define their respective goals and functions, build knowledge and data storage, build risk assessment models or environmental impact analysis models, build decision-making mechanisms, and design communication interfaces. The game mechanism between insurance companies and reinsurance companies includes: Under stable market conditions, insurance companies and reinsurance companies optimize contracts through Stackelberg reinforcement learning. The reinsurance company first formulates reinsurance contract parameters with the goal of maximizing profits. These reinsurance contract parameters include reinsurance rates, reinsurance ratios, and deductibles. After the reinsurer sets the contract parameters, the insurance company selects the optimal reinsurance purchase strategy based on its own capital adequacy ratio and market risk. In the event of severe market volatility or catastrophic events, insurance and reinsurance companies choose to cooperate and share profits according to the Shapley value distribution mechanism.

2. The method for optimizing a catastrophe reinsurance contract according to claim 1, characterized in that, Reinsurance contract parameter optimization strategies include: reinsurance ratio optimization, reinsurance rate optimization, and deductible optimization; The reinsurance ratio is optimized according to the following formula. : In the formula, For the optimal reinsurance ratio, This indicates penalties for failing to meet regulatory requirements regarding capital adequacy ratios. The penalty coefficient is used to balance profit and the risk of SCR violation. The expected profit of the product line; Optimize reinsurance rates using the following formula: In the formula, P* is the optimal reinsurance rate, E[L|L>d] is the expected loss after exceeding the deductible d, and ζ is the profit loading factor of the reinsurance company; Optimize the deductible using the following formula: In the formula, To be the optimal deductible, It is the degree of risk exposure. It is a risk adjustment parameter.

3. The method for optimizing a catastrophe reinsurance contract according to claim 1, characterized in that, The contract parameters for multi-agent cooperative optimization include: For insurance company agents, the reinsurance ratio is adjusted based on capital adequacy ratio and market conditions; For reinsurance companies' intelligent agents, reinsurance rates and deductibles are adjusted according to market competition. For regulatory compliance agents, provide real-time compliance checks to ensure that capital adequacy ratios meet regulatory requirements; For risk assessment agents, market risk forecasts are provided, which affect decisions regarding reinsurance ratios and deductibles; For intelligent agents targeting the external environment, monitoring market economic conditions influences rate adjustments.

4. The method for optimizing a catastrophe reinsurance contract according to claim 1, characterized in that, Multi-agent reinforcement learning methods include: When making decisions, each agent defines a state vector based on the market environment, contract terms, and historical data. For insurance company agents, set reinsurance ratios and whether to accept reinsurance contracts; for reinsurance company agents, set reinsurance rates and deductibles; for risk assessment agents, determine whether to renegotiate reinsurance contract terms; for regulatory compliance agents, determine whether to change the minimum solvency adequacy ratio; for external environment agents, determine whether to change risk exposure. Design an agent reward function and adopt an Actor-Critic architecture to optimize contract parameters. The Actor network is used to generate the optimal contract policy, and the Critic network is used to evaluate the performance of the current policy in long-term returns and adjust the policy accordingly. Initialize the neural network for each agent, the Replay Buffer for storing the agents' historical interaction data, and set the learning rate and discount factor; Each agent interacts with the environment based on the current policy, obtains the state, selects actions, receives immediate rewards and the next state, and stores the interaction results in the Replay Buffer; Training is performed by randomly sampling from the Replay Buffer, calculating the Q-value update for each agent, updating the agent's policy using the deterministic policy gradient, and estimating the Q-value using the target network. The training process terminates when the agent's policy converges after multiple training epochs and the long-term reward of each agent reaches its optimum.

5. The method for optimizing a catastrophe reinsurance contract according to claim 4, characterized in that, The strategy is updated according to the following formula: In the formula, It's the learning rate. It is a state-action value function. It is about gradient, It is about The gradient.

6. An optimization system for catastrophe reinsurance contracts applied to the optimization method of catastrophe reinsurance contracts according to any one of claims 1-5, characterized in that, include: Building blocks are used to construct multi-agent systems; The design unit is used to design a dynamically optimized reinsurance contract mechanism by employing a mathematical optimization model and combining the core clauses in the reinsurance contract to optimize the reinsurance contract terms. The optimization of the reinsurance contract mechanism includes a reinsurance contract parameter optimization strategy and multi-agent collaborative optimization of contract parameters. The optimization unit is used to design the cooperation and game mechanism between the agents and to dynamically optimize the game strategy using a multi-agent reinforcement learning method to adjust the contract terms in real time. The cooperation and game mechanism between the agents includes the game mechanism between the insurance company and the reinsurance company, as well as the agent coordination mechanism.

7. An electronic 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 program, it implements the steps of the optimization method for the catastrophe reinsurance contract as described in any one of claims 1 to 5.

8. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the optimization method for a catastrophe reinsurance contract as described in any one of claims 1 to 5.

Images