Navigation industry energy transformation dynamic prediction method based on multilevel modeling

Abstract

The invention provides a multi-level modeling-based shipping industry energy transformation dynamic prediction method. The method comprises the steps of obtaining shipping industry data; constructing an agent model from a microscopic level; constructing an evolutionary game model from the mesoscopic level; the method comprises the following steps: constructing a Lotka-Volterra model from a macroscopic level; establishing a bidirectional feedback mechanism; and outputting long-term market shares and competition patterns of different fuel technologies. According to the shipping industry energy transformation dynamic prediction method based on multi-level modeling, the advantages of three models are integrated, and the shipping industry energy transformation dynamic process can be comprehensively described on multiple levels.

Description

Technical Field

[0001] The present invention relates to the technical field of energy conservation and environmental protection, and in particular to a dynamic prediction method for energy transformation in the shipping industry based on multi-level modeling. Background Art

[0002] In view of the complexity of energy transformation in the shipping industry, existing studies usually adopt single-level modeling methods (such as agent models, evolutionary game models and Lotka-Volterra models), but these methods have significant limitations in practical applications and are difficult to fully analyze the dynamic process of energy transformation. First, the agent model (ABM) can capture individual decision-making mechanisms and local interaction effects by simulating the behavior of heterogeneous individuals (such as shipping companies), but its shortcomings are the lack of analytical ability for group strategy dynamics and the inability to reveal the evolutionary laws and long-term stability of strategy distribution from a holistic level. In addition, when the number of agents is huge or the decision rules are complex, the computational cost increases significantly, making it difficult to make large-scale or long-term predictions. Secondly, although the evolutionary game model (EGT) has advantages in the dynamic evolution and long-term stability analysis of strategy distribution, it assumes individual homogeneity and is difficult to reflect the heterogeneous characteristics of shipping companies in decision-making behavior. It also ignores local interaction mechanisms (such as competition and cooperation relationships), which limits the ability to characterize complex market behaviors. At the same time, since the model input parameters are usually static, the policy sensitivity analysis ability of EGT is limited, and it is difficult to dynamically reflect the impact of policy changes on the system. Finally, the Lotka-Volterra model (LV), as a classic competitive dynamics model, can be used to describe the long-term competitive relationship between fuel technologies, but it lacks a micro-behavior mechanism and it is difficult to explain how micro-individual behavior drives macro-dynamic changes. In addition, the LV model requires a fixed initial strategy distribution and parameters. In the energy transition of the shipping industry, the strategy distribution and parameters have significant time dynamics, and the model is difficult to adapt to this feature. In summary, the existing single model method has obvious shortcomings in capturing the multi-level complex dynamics of the energy transition in the shipping industry. There is an urgent need for an innovative framework that can integrate multi-level modeling methods and dynamically couple micro-behavior and macro-competition to fully reveal the inherent laws of energy transition.

[0003] The limitations of the existing single model method make it difficult to meet the key needs of the research on energy transition in the shipping industry. First, energy transition is a complex process involving micro-individual behavior, group strategy dynamics and macro-competitive landscape, which requires multi-level analysis capabilities to fully reveal the dynamic evolution from individual decision-making to market competition. However, existing methods often only focus on one level and cannot simultaneously characterize individual behavior, dynamic changes in strategy distribution and technological competition landscape. Second, the two-way interaction between micro-behavior and macro-dynamics is the core feature of energy transition. It is urgent to establish dynamic coupling capabilities to reveal how individual behavior affects the macro market structure and analyze how macro trends react to individual decisions. However, the existing model lacks this two-way feedback mechanism and cannot effectively deal with this complex relationship. In addition, policy sensitivity analysis capabilities are also an important need for current research. Carbon taxes, fuel subsidies and other policies have a significant impact on shipping companies' fuel choices and market competition patterns. A modeling framework that can dynamically evaluate the short-term and long-term effects of policies is needed to assist policy formulation and effect evaluation. Finally, accurate prediction of the long-term market share of different fuel technologies is an important basis for planning, but the existing methods are insufficient in modeling long-term trends and cannot provide reliable support for policymakers and industry participants. Therefore, research urgently needs to break through the limitations of existing models and develop a comprehensive modeling method with multi-level dynamic analysis, two-way feedback coupling, policy sensitivity assessment and long-term trend forecasting capabilities to scientifically respond to the complexity of energy transformation in the shipping industry.

[0004] The invention patent application with application number CN 115115262A discloses a decision-making method for energy transformation of high-energy-consuming enterprises based on carbon emission prediction under multiple scenarios, including the following steps: designing future carbon emission scenarios from the two dimensions of economic development and energy consumption; forming at least 8 different scenario combinations according to the design situation; identifying the influencing factors of carbon emission of high-energy-consuming enterprises; constructing a preliminary STIRPAT model for carbon emission prediction of high-energy-consuming enterprises; performing collinearity test on each influencing factor through a regression model; correcting the STIRPAT model for carbon emission prediction of high-energy-consuming enterprises according to the collinearity test results; predicting carbon emission indicators of high-energy-consuming enterprises under different scenarios; calculating economic indicators of energy transformation under different scenarios; weighting carbon emission indicators and economic indicators through the entropy weight method; calculating energy transformation evaluation scores under different scenarios; and making energy transformation decisions for high-energy-consuming enterprises. The disadvantage of this method is that it uses a single model method to make energy transformation decisions for high-energy-consuming enterprises based on carbon emission prediction under multiple scenarios, which has significant limitations in practical applications and is difficult to meet the key needs of energy transformation research in the shipping industry.

[0005] The invention patent application with the application number CN 108364135 A discloses an energy transition decision-making support method based on the interaction simulation of technology-economic-real participants-computer agents. The transformation decisions of real participants are added to the quantitative energy system technology-economic model. At each simulation step, the decision-making behaviors of real participants and computer agents are used as the input quantities of the objective technology-economic model. The decisions in the next stage are adjusted based on the simulation results driven by the behavioral decisions in the previous stage, so as to reflect the interactive influence of technology-economic-multi-party game behaviors in the time trajectory of the simulation results. The disadvantage of this method is that it uses the method of simulation experiments for energy transition decision-making, with insufficient modeling ability for long-term trends and difficulty in providing reliable support for policymakers and industry participants. Summary of the Invention

[0006] In order to solve the above technical problems, a dynamic prediction method for the energy transition of the shipping industry based on multi-level modeling is proposed in the present invention. By integrating the advantages of three models, the present invention can comprehensively depict the dynamic process of the energy transition of the shipping industry at multiple levels.

[0007] The object of the present invention is to provide a dynamic prediction method for the energy transition of the shipping industry based on multi-level modeling, including obtaining shipping industry data, and further including the following steps: Step 1: Construct an agent model from the micro level; Step 2: Construct an evolutionary game model from the meso level; Step 3: Construct a Lotka-Volterra model from the macro level; Step 4: Establish a two-way feedback mechanism; Step 5: Output the long-term market shares and competition patterns of different fuel technologies.

[0008] Preferably, the shipping industry data includes shipping company attributes, fuel technology parameters, and dynamic policy variables.

[0009] In any of the above solutions, preferably, the shipping company attributes include at least one of ship size, current fuel selection, and capital capacity.

[0010] In any of the above solutions, preferably, the fuel technology parameters include at least one of fuel cost, carbon emission level, technology conversion cost, and market price.

[0011] In any of the above solutions, preferably, the dynamic policy variables include at least one of carbon tax, fuel subsidy, and R & D incentive.

[0012] In any of the above solutions, preferably, each agent represents a shipping company, and the key attributes of the agent include: 1) The number and attributes of ships; 2) Current fuel selection; 3) Financial capacity; 4) Technology acceptance ability.

[0013] Preferably in any of the above solutions, step 1 includes the following sub-steps: Step 11: Use the shipping industry data as the input of the agent model; Step 12: Define the revenue function for each agent; Step 13: Output the strategy distribution of fuel selection and the revenue function of each strategy U k 。

[0014] Preferably in any of the above solutions, the calculation formula of the revenue function is U k = R k - C k + S k - T k Wherein, U k is the revenue of strategy k of, R k is the market revenue, C k is the fuel cost, S k is the policy subsidy, T k is the technology conversion cost.

[0015] Preferably in any of the above solutions, use a rule-based decision-making mechanism and / or a probability-based decision-making mechanism to simulate the strategy selection process of the agent.

[0016] Preferably in any of the above solutions, step 2 includes the following sub-steps: Step 21: Use the strategy distribution and the revenue function U k as the input of the agent model; Step 22: Use the replicator dynamics equation to simulate the dynamic change of the strategy distribution; Step 23: By analyzing the stable points of the equation, judge whether there is an evolutionarily stable strategy ESS; Step 24: Output the dynamic change rate of the strategy distribution and the proportion of the strategy in the stable state.

[0017] Preferably, in any of the above solutions, the calculation formula for the dynamic change of the strategy distribution is

[0018]

[0019]

[0020] where is the strategy k at time t growth rate, x k is the proportion of the strategy k , is the average group income.

[0021] Preferably, in any of the above solutions, step 3 includes the following sub-steps: Step 31: Using the strategy distribution dynamics output by the evolutionary game model, extract the preliminary strategy proportion x k (0), strategy growth rate r k and strategy competition coefficient α ij , as the input of the Lotka-Volterra model; Step 32: Through the Lotka-Volterra equation, simulate the long-term market competition relationship of different fuel technologies; Step 33: Output the long-term market share and competition pattern of at least one of fuel oil, LNG and methanol.

[0022] Preferably, in any of the above solutions, the strategy growth rate r k calculation formula is

[0023]

[0024] Preferably, in any of the above solutions, the strategy competition coefficient α ij calculation formula is

[0025]

[0026] where is the growth rate change of the strategy i caused by the strategy i , is the growth rate change of the strategy j caused by the strategy j .

[0027] Preferably, in any of the above solutions, the due to the strategyi Resulting strategy i The calculation formula for the growth rate change of is

[0028]

[0029] Wherein, r i is the growth rate of strategy i of x i is the market share of strategy i of x j is the market share of strategy j of

[0030] Preferably, in any of the above solutions, step 4 includes the following sub-steps: Step 41: Establish a feedback from the micro level to the macro level; Step 42: Establish a feedback from the macro level to the micro level; Step 43: Run in a cyclic iteration until the strategy distribution and the market share converge, forming a dynamic coupling modeling framework from micro to macro.

[0031] Preferably, in any of the above solutions, step 41 includes using the output of the agent model as the input of the evolutionary game model to further analyze the strategy dynamics.

[0032] Preferably, in any of the above solutions, step 42 includes feeding back the long-term prediction result of the Lotka-Volterra model to the agent model to dynamically adjust the payoff function.

[0033] The present invention proposes a dynamic prediction method for the energy transformation of the shipping industry based on multi-level modeling, which can comprehensively solve the problems of technical path selection, market competition relationship analysis and policy effectiveness evaluation, and provide scientific and practical tool support for shipping energy transformation policy research and enterprise low-carbon planning.

[0034] The Lotka-Volterra model is a classic mathematical model that describes the dynamic relationship between predator and prey populations in an ecosystem. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flowchart of a preferred embodiment of the dynamic prediction method for the energy transformation of the shipping industry based on multi-level modeling according to the present invention.

[0036] Figure 2 is a flowchart of another preferred embodiment of the dynamic prediction method for the energy transformation of the shipping industry based on multi-level modeling according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0037] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0038] Embodiment 1 As Figure 1 shown, a dynamic prediction method for the energy transformation of the shipping industry based on multi-level modeling, which performs step 100 to obtain shipping industry data. The shipping industry data includes shipping company attributes, fuel technical parameters, and dynamic policy variables. The shipping company attributes include at least one of ship size, current fuel selection, and capital capacity. The fuel technical parameters include at least one of fuel cost, carbon emission level, technology conversion cost, and market price.

[0039] The dynamic policy variables include at least one of carbon tax, fuel subsidy, and R & D incentive.

[0040] Perform step 110 to construct an agent model at the micro level, including the following sub-steps: Perform step 111 to use the shipping industry data as the input of the agent model.

[0041] Perform step 112 to define a revenue function for each agent. The calculation formula of the revenue function is U k = R k - C k + S k - T k Wherein, U k is the revenue of strategy k , R k is the market revenue, C k is the fuel cost, S k is the policy subsidy, T k is the technology conversion cost.

[0042] Perform step 113 to output the strategy distribution of fuel selection and the revenue function of each strategy U k .

[0043] Each agent represents a shipping company. The key attributes of the agent include: 1) The number and attributes of ships; 2) The current fuel selection; 3) Financial capability; 4) Technology acceptance.

[0044] The strategy selection process of the agent is simulated using a rule-based decision-making mechanism and / or a probability-based decision-making mechanism.

[0045] Execute step 120 to construct an evolutionary game model from a meso-level, including the following sub-steps: Execute step 121 to convert the strategy distribution and the profit function U k As input to the proxy model.

[0046] Execute step 122, use the replication dynamic equation to simulate the dynamic change of the strategy distribution, the calculation formula of the dynamic change of the strategy distribution is:

[0047]

[0048]

[0049] in, For strategy k In time t The growth rate, x k For strategy k The proportion of is the average income of the group.

[0050] Execute step 123 to determine whether there is an evolutionary stable strategy ESS by analyzing the stable point of the equation.

[0051] Execute step 124 to output the dynamic change rate of the strategy distribution and the strategy proportion in the stable state.

[0052] Execute step 130 to construct the Lotka-Volterra model from a macro level, wherein step 3 includes the following sub-steps: Execute step 131, use the strategy distribution dynamics output by the evolutionary game model to extract the initial strategy proportion x k (0), Strategy growth rate r k and the strategic competition coefficient α ij , as the input of the Lotka-Volterra model, the strategy growth rate r k The calculation formula is

[0053]

[0054] The strategy competition coefficient αij The calculation formula is

[0055]

[0056] in, Because of the strategy i The resulting strategy i The change in growth rate, Because of the strategy j The resulting strategy j changes in growth rates.

[0057] Execute step 132 to simulate the long-term market competition relationship of different fuel technologies through the Lotka-Volterra equation. i The resulting strategy i The formula for calculating the change in growth rate is

[0058]

[0059] in, r i For strategy i The growth rate, x i For strategy i market share, x j For strategy j market share.

[0060] Execute step 133 to output the long-term market share and competition pattern of at least one fuel among fuel oil, LNG and methanol.

[0061] Execute step 140 to establish a two-way feedback mechanism, wherein step 4 includes the following sub-steps: Execute step 141 to establish feedback from the micro level to the macro level, including using the output of the agent model as the input of the evolutionary game model to further analyze the strategy dynamics.

[0062] Execute step 142 to establish feedback from the macro level to the micro level, including feeding back the long-term prediction results of the Lotka-Volterra model to the proxy model to dynamically adjust the benefit function Execute step 143 and iterate the loop until the strategy distribution and market share converge, forming a dynamic coupling modeling framework from micro to macro.

[0063] Step 150 is executed to output the long-term market share and competition pattern of different fuel technologies.

[0064] Example 2 In order to solve the shortcomings of existing research in multi-level analysis, dynamic coupling, policy sensitivity analysis and long-term prediction capabilities, this paper proposes a multi-level modeling method based on the technical framework of "agent model-evolutionary game model-Lotka-Volterra model". By integrating the advantages of the three models, the present invention can comprehensively characterize the dynamic process of energy transformation in the shipping industry at multiple levels. At the individual level, the agent model (ABM) is used to simulate the heterogeneous decision-making behavior of shipping companies under policy and market changes, revealing the decision-making mechanism of micro-individuals; at the group level, the dynamic evolution process of strategy distribution is studied through the evolutionary game model (EGT), and whether there is an evolutionary stable strategy (ESS) in the group strategy; at the macro level, the Lotka-Volterra model (LV) is used to predict the long-term market share and competitive landscape of different fuel technologies, revealing the macro dynamic trend. At the same time, through the two-way data coupling mechanism, the present invention realizes dynamic linkage analysis from micro to macro, revealing how individual behavior affects the macro market structure, and how macro trends react to individual decisions.

[0065] In practical applications, the present invention can better support the research and optimization of shipping energy transformation policies, and provide a scientific basis for policy makers by dynamically evaluating the short-term and long-term impacts of policies such as carbon taxes and fuel subsidies on fuel selection and market competition patterns. In addition, the present invention can also meet the needs of shipping companies in formulating low-carbon development plans, and provide decision support for companies to evaluate technology investment risks, optimize fuel selection strategies, and formulate long-term competitive strategies by simulating market dynamics under different fuel technologies and policy scenarios. Therefore, this multi-level modeling method not only has important theoretical value, but also has broad application prospects in policy research and corporate practice of shipping energy transformation.

[0066] In view of the complexity of energy transformation in the shipping industry and the shortcomings of existing research, this paper proposes a multi-level modeling method from micro to macro and dynamic coupling, which can comprehensively solve the problems of technology path selection, market competition relationship analysis and policy effectiveness evaluation, and provide scientific and practical tool support for shipping energy transformation policy research and corporate low-carbon planning.

[0067] The present invention can comprehensively analyze the energy transformation process of the shipping industry from micro individual behavior to macro competition pattern. The technical solution includes the following core steps: 1. Micro level: Construction of agent model Define the proxy properties: Each agent represents a shipping line and sets its key attributes, including: 1. Number and type of ships; 2. Current fuel selection; 3. Financial capacity (such as capital reserves); 4. Technology acceptance (such as adaptability to new fuels).

[0068] Construct the profit function: The agent's payoff function takes into account the following dynamic variables: 1. Fuel cost: including base fuel price and carbon tax cost; 2. Policy subsidies: incentives for low-carbon or zero-carbon fuels; 3. Market revenue: adjusted according to the market acceptance of fuel technology; 4. Technology conversion cost: the input required to switch from one fuel technology to another.

[0069] U k = R k - C k + S k - T k in, U k For strategy k of income; R k For market income; C k For fuel costs; S k To provide policy subsidies; T k The cost of technology conversion.

[0070] Set up decision rules: 1. Use rule-based decision-making mechanisms (such as profit maximization) or probability-based decision-making mechanisms (such as the Logit model) to simulate the agent's strategy selection process; 2. Output: The proportion of each fuel strategy (strategy distribution).

[0071] 2. Meso-level: Construction of evolutionary game model

[0072] Input the output of the proxy model: The strategy distribution of the agent model (such as the proportion of fuel oil, LNG and methanol) is used as the initial condition of the evolutionary game model.

[0073] Construct the dynamic evolution equation: Based on the replication dynamic equation, simulate the dynamic evolution of strategy distribution:

[0074]

[0075]

[0076] in, x k For strategy k The proportion of U k For strategy k The income of is; is the average income of the group: Analyzing Evolutionarily Stable Strategies (ESS): By analyzing the stable points of the equation, we can determine whether there is an evolutionarily stable strategy (ESS). ESS represents the long-term stable state of the group strategy distribution under current policy and market conditions.

[0077] Output strategy dynamic trend: The evolutionary game model outputs the dynamic change rate and long-term stable proportion of each strategy.

[0078] 3. Macro level: Construction of Lotka-Volterra model Extract key parameters: Using the strategy distribution dynamics output by the evolutionary game model, the following parameters are extracted: Initial strategy ratio x k (0) Strategy growth rate

[0079] Competition coefficient between strategies α ij Constructing the Lotka-Volterra model: Describe the long-term competition between different fuel technologies:

[0080]

[0081] in, x i For strategy i market share; r i For strategy i growth rate; αij For strategy j Strategy i intensity of competition.

[0082] Forecasted long-term market share: By simulating the Lotka-Volterra model, the long-term market share and competitive landscape of different fuel technologies are output.

[0083] 4. Two-way feedback mechanism Feedback from micro to macro: The strategy distribution dynamics of the agent model is used as the input of the evolutionary game model to further derive the long-term changes of the strategy.

[0084] Feedback from macro to micro: The long-term prediction results of the Lotka-Volterra model are fed back to the agent model to dynamically adjust the agent's profit function (such as updating fuel costs or policy subsidies).

[0085] Loop iteration: Through multiple rounds of iterations, a dynamic coupling modeling framework from micro to macro is formed to achieve a comprehensive analysis of the energy transformation of the shipping industry.

[0086] The technical solution of the present invention has the following significant advantages: Multi-level dynamic analysis capabilities: By integrating the agent model, evolutionary game model and Lotka-Volterra model, a full range of dynamic analysis is achieved from micro behavior to macro competition.

[0087] Bidirectional dynamic coupling mechanism: A two-way feedback mechanism between micro-individual decisions and macro-market patterns has been established, which can dynamically adjust the model to adapt to policy and market changes.

[0088] Policy sensitivity analysis: Able to dynamically evaluate the short-term and long-term impacts of different policies (such as carbon tax, fuel subsidies) on shipping companies' fuel choices and market competition landscape.

[0089] Long-term trend forecasting capabilities: The Lotka-Volterra model can be used to accurately predict the long-term market share and competitive landscape of different fuel technologies, providing a scientific basis for policy making and industry planning.

[0090] Example 3 The present invention proposes a dynamic prediction method for energy transformation in the shipping industry based on multi-level modeling. Through the three-layer framework of agent model, evolutionary game model and Lotka-Volterra model, the dynamic process of fuel selection of shipping companies, the evolution law of strategy distribution and the long-term competition pattern of different fuel technologies are gradually simulated.

[0091] The system consists of the following key modules: Agent model module: simulates the fuel selection behavior of shipping companies under different policy scenarios, and outputs strategy distribution and profit data.

[0092] Evolutionary game model module: simulates the dynamic evolution of strategy distribution and analyzes long-term stable strategies (ESS).

[0093] Lotka-Volterra model module: simulates the long-term market share and competitive relationship of different fuel technologies.

[0094] Bidirectional feedback mechanism module: realizes data transmission and dynamic iteration between models.

[0095] like Figure 2 The following are the specific implementation steps: Step 1: Proxy Model Module Input data: 1. Shipping company attributes: including vessel size, current fuel options, capital capacity, etc.; 2. Fuel technical parameters: including fuel cost, carbon emission level, technology conversion cost, market price, etc.; 3. Policy scenarios: including dynamic policy variables such as carbon tax, fuel subsidies, and R&D incentives.

[0096] Model building and running: Define the revenue function for each shipping company: U k = R k - C k + S k - T k U k :Strategy k of income; R k : Market revenue; C k : fuel cost; S k :Policy subsidies; T k : Technology switching cost.

[0097] The fuel selection behavior of agents is simulated according to the profit maximization principle or the Logit model.

[0098] Output: 1. Distribution of fuel selection strategies (e.g., the proportion of fuel oil, LNG, and methanol); 2. Returns of each strategy U k .

[0099] Step 2: Evolutionary Game Model Module Input data: Strategy distribution: The proportion of fuels such as fuel oil, LNG and methanol output by the proxy model.

[0100] Strategy return: the returns of each strategy output by the proxy model U k .

[0101] Model building and running: The dynamic changes of strategy distribution are simulated using the replication dynamic equation:

[0102]

[0103] x k :Strategy k The proportion of :The average income of the group, calculated by:

[0104] Output: 1. The rate of change of the policy distribution; 2. The proportion of strategies in the stable state (evolutionarily stable strategy, ESS).

[0105] Step 3: Lotka-Volterra Model Module Input data: Initial strategy ratio x k (0): Strategy distribution output of evolutionary game model.

[0106] Strategy growth rate r k :Calculated according to the evolutionary game model:

[0107]

[0108] Strategic Competition Coefficient α ij : Derived based on the profit function, or approximated by the following formula:

[0109]

[0110] Model building and running: The long-term market competition relationship between different fuel technologies is simulated by the Lotka-Volterra equation:

[0111]

[0112] Output: The long-term market share and competition landscape of fuels such as fuel oil, LNG, and methanol.

[0113] Step 4: Bidirectional Feedback Mechanism Module Feedback from micro to macro: The output of the agent model (strategy distribution and payoff data) is used as the input of the evolutionary game model to further analyze the strategy dynamics.

[0114] Feedback from macro to micro: The long-term prediction results of the Lotka-Volterra model are fed back to the agent model to dynamically adjust the benefit function (such as fuel costs and policy subsidies).

[0115] Loop iteration: Multiple rounds of iterations are run until the strategy distribution and market share converge.

[0116] In order to better understand the present invention, the above is described in detail in conjunction with the specific embodiments of the present invention, but it is not intended to limit the present invention. Any simple modifications made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention. Each embodiment in this specification focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referenced to each other. For the system embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

Claims

1. A dynamic prediction method for energy transformation in the shipping industry based on multi-level modeling, comprising obtaining shipping industry data, and further comprising the following steps: Step 1: Construct an agent model at the micro level, including the following sub-steps: Step 11: Using the shipping industry data as input to the agent model; Step 12: Define the payoff function for each agent; Step 13: Output the strategy distribution of fuel selection and the profit function of each strategy U k ; Step 2: Construct an evolutionary game model from the meso-level, including the following sub-steps: Step 21: Distribute the strategy and the profit function U k as input to the proxy model; Step 22: Use the replication dynamic equation to simulate the dynamic changes of strategy distribution; Step 23: By analyzing the stable points of the equation, determine whether there is an evolutionary stable strategy ESS; Step 24: Output the dynamic change rate of the strategy distribution and the strategy proportion in the stable state; Step 3: Construct the Lotka-Volterra model from a macro level, including the following sub-steps: Step 31: Use the strategy distribution dynamics output by the evolutionary game model to extract the initial strategy proportions x k (0), Strategy growth rate r k and the strategic competition coefficient α ij , as input to the Lotka-Volterra model; Step 32: Use the Lotka-Volterra equation to simulate the long-term market competition relationship between different fuel technologies; Step 33: Output the long-term market share and competitive landscape of at least one of fuels including fuel oil, LNG and methanol; Step 4: Establish a two-way feedback mechanism, including the following sub-steps: Step 41: Establishing feedback from the micro level to the macro level; Step 42: Establishing feedback from the macro level to the micro level; Step 43: Run the loop iteratively until the strategy distribution and market share converge, forming a dynamic coupling modeling framework from micro to macro; Step 5: Output the long-term market share and competitive landscape of different fuel technologies.

2. The method for dynamic prediction of energy transformation in the shipping industry based on multi-level modeling as claimed in claim 1, characterized in that: The calculation formula of the profit function is: U k = R k - C k + S k - T k in, U k For strategy k of income, R k For market income, C k For fuel costs, S k For policy subsidies, T k The cost of technology conversion.

3. The method for dynamic prediction of energy transformation in the shipping industry based on multi-level modeling as claimed in claim 2, characterized in that: The strategy selection process of the agent is simulated using a rule-based decision-making mechanism and / or a probability-based decision-making mechanism.

4. The method for dynamic prediction of energy transformation in the shipping industry based on multi-level modeling as claimed in claim 3 is characterized in that: The calculation formula for the dynamic change of the strategy distribution is: ， ， in, For strategy k In time t The growth rate, x k For strategy k The proportion of is the average income of the group.

5. The method for dynamic prediction of energy transformation in the shipping industry based on multi-level modeling as claimed in claim 4, characterized in that: The growth rate of the strategy r k The calculation formula is 。 6. The method for dynamic prediction of energy transformation in the shipping industry based on multi-level modeling as claimed in claim 5, characterized in that: The strategy competition coefficient α ij The calculation formula is ， in, Because of the strategy i The resulting strategy i The change in growth rate, Because of the strategy j The resulting strategy j changes in growth rate.

7. The method for dynamic prediction of energy transformation in the shipping industry based on multi-level modeling as claimed in claim 6, characterized in that: The strategy i The resulting strategy i The formula for calculating the change in growth rate is ， in, r i For strategy i The growth rate, x i For strategy i market share, x j For strategy j market share.

Images