Optimization method for project investment decision

Through multi-dimensional data integration and machine learning model analysis, combined with real-time market monitoring and multi-objective optimization, Monte Carlo simulation and Bayesian network were introduced, which solved the problem of insufficient dependence on historical data and risk assessment in the existing technology, and achieved more Scientific and accurate project investment decisions, adapt to market changes and reduce risks.

CN119048153AInactive Publication Date: 2024-11-29CCCC SHEC DONGMENG ENG CO LTD
View PDF 0 Cites 8 Cited by

Patent Information

Application Number
CN202411128312.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-16
Publication Date
2024-11-29
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing project investment decision-making methods rely too much on historical data, are difficult to capture market changes in a timely manner, ignore multi-source data integration, lack dynamic adjustment and adaptive optimization capabilities, and insufficient risk assessment, resulting in decision-making lag and investment losses.

Method used

Using multi-dimensional data integration, analyzing project characteristics through machine learning models, building a real-time market dynamic monitoring system, combining multi-objective optimization algorithms to prioritize investments, and introducing Monte Carlo simulation and Bayesian network for risk assessment, and dynamically adjusting investments combination.

Benefits of technology

It improves the scientificity and accuracy of investment decisions, adapts to the rapidly changing market environment, enhances risk management capabilities, and ensures that the investment portfolio maintains its best state in market fluctuations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119048153A_ABST
    Figure CN119048153A_ABST
Patent Text Reader

Abstract

The invention relates to a project investment decision optimization method. The method comprises the following steps: firstly, collecting and integrating multi-dimensional data related to a project, including financial data of the project, project team background information, market demand prediction, industry trend analysis and competitiveness evaluation of the project; a real-time market dynamic monitoring system is constructed, and potential influences of market changes on projects are analyzed by integrating market data including stock market fluctuation, industry news and policy changes; based on the project scoring matrix and the project adaptability analysis result, performing investment priority ranking on the projects by using a multi-objective optimization algorithm; on the basis of investor resource limitation and an investment priority ranking result, a mixed integer programming or dynamic programming algorithm is adopted, a resource optimization configuration model is established, and under the condition that resources are limited, a resource allocation strategy is optimized, and a project combination with potential is selected for investment; and introducing Monte Carlo simulation or a Bayesian network, and evaluating uncertainty and potential risks in the project investment decision.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of project investment, and in particular to an optimization method for project investment decision-making. Background Art

[0002] Although the current project investment decision optimization method has certain advantages, its shortcomings and disadvantages are gradually emerging in the face of a complex and changing market environment, limiting investors' decision-making ability and the realization of maximizing investment returns in a highly dynamic market. For example, the project investment decision method and system disclosed in patent CN115809928A, first of all, this method relies too much on historical data for modeling and prediction. The investment decision system collects data of multiple preset dimensions and performs feature engineering based on these data to obtain project features. The core of this method is to rely on an integrated learning model driven by historical data to predict the investable probability of a project. This method may be more effective in a stable market environment because historical data can often reflect market trends and laws. However, the market environment is dynamically changing, especially changes in technology, policies, economic conditions, etc. will have a significant impact on the market. Relying solely on historical data for modeling may not be able to capture the latest changes and emergencies in the market in a timely manner, resulting in delayed decision-making and missed market opportunities. In addition, historical data often contains outdated information, which may no longer be applicable in the new market environment, resulting in deviations in the prediction model. For emerging markets or rapidly developing industries, historical data has limited reference value, and these markets are important areas for investors to pursue high returns. Therefore, over-reliance on historical data for decision-making may limit investors' decision-making ability in these areas. Secondly, this method has certain limitations in the process of project data collection and processing.

[0003] The project investment decision-making system of CN115809928A emphasizes the collection of data related to investment decisions through preset dimensions, and performs feature engineering based on these data. Although this method can ensure the consistency of data collection, it ignores the potential of multi-source data integration. In reality, the factors that affect project investment decisions are often multifaceted, involving multiple dimensions such as financial market data, policy changes, industry news, and socio-economic changes. These data sources vary, the data structure is complex, and the data update frequency is different. It is difficult to fully cover all information that may affect investment decisions by simply relying on preset dimensions. Third, although the investment decision-making system in CN115809928A can predict investment probabilities through an integrated learning model, it lacks the ability to dynamically adjust and adaptively optimize. In this system, project features are input into a model based on integrated learning, and the system makes decisions based on the investable probability output by the model. This static decision-making process ignores the dynamics of the market environment and the constant changes in the project itself. In reality, factors such as market conditions, policy environment, and technological progress are constantly changing, and the investment value of the project fluctuates accordingly. Fourth, the system in CN115809928A lacks a comprehensive risk assessment mechanism. Risk management is a crucial part in project investment decisions. However, the system only predicts the probability of investment through an integrated learning model, and fails to fully consider the risk distribution of projects in different market environments and potential market fluctuations. Modern investment decision-making methods usually introduce multi-model risk assessment tools, such as Monte Carlo simulation and Bayesian networks, to comprehensively analyze the performance and risk level of projects in different market scenarios. Through such risk assessment, investors can better identify high-risk projects and adopt corresponding reduction or exit strategies to reduce the risk of the overall investment portfolio. In addition, the structured risk assessment of the Bayesian network can also dynamically update the dependency between projects and market factors, further improving the accuracy and security of investment decisions. In contrast, the system of CN115809928A is relatively weak in risk management, which may cause investors to be unable to make effective adjustments in time when facing high-risk projects, thereby increasing the possibility of investment losses. Summary of the invention

[0004] The purpose of the present invention is to provide an optimization method for project investment decision-making, thereby solving some of the drawbacks and shortcomings pointed out in the background technology.

[0005] The present invention solves the above-mentioned technical problems by adopting the following technical solutions, including:

[0006] S1. Collect and integrate multi-dimensional data related to the project, including financial data of the project, background information of the project team, market demand forecast, industry trend analysis and competitiveness assessment of the project;

[0007] S1.1. Analyze project characteristics through machine learning models, extract key indicators, and conduct multi-dimensional quantitative modeling for each project to generate a project scoring matrix;

[0008] S2. Build a real-time market dynamics monitoring system to analyze the potential impact of market changes on projects by integrating market data, including stock market fluctuations, industry news and policy changes;

[0009] S2.1. Develop a project adaptability analysis module that combines market dynamics data with project characteristic data to assess the performance and potential of each project under current market conditions;

[0010] S3. Based on the project scoring matrix and the project adaptability analysis results, the investment priorities of the projects are sorted using a multi-objective optimization algorithm;

[0011] S3.1. Update the investment priorities of the projects in real time according to market changes and project progress;

[0012] S4. Based on the resource constraints of investors, the investment priority ranking results are used to establish a resource optimization allocation model using mixed integer programming or dynamic programming algorithms. Under the condition of limited resources, the resource allocation strategy is optimized and a potential project portfolio is selected for investment;

[0013] S5. Introduce Monte Carlo simulation or Bayesian network to evaluate the uncertainty and potential risks in the investment decision of the project;

[0014] S5.1. Adjust the investment portfolio according to the risk assessment results.

[0015] Furthermore, generating a project rating matrix comprises the following steps:

[0016] S1. Use machine learning models to extract and select project features, and build an adaptive learning mechanism; define the function Ψ(x) to represent the process of extracting key features from a large number of data sets, where x is the feature vector of the project, and the form of Ψ(x) is as follows:

[0017]

[0018] Among them, α i is the weight parameter obtained by adaptive learning, f i (x) represents the feature extraction function, g(t) is the time-dependent feature weight adjustment function;

[0019] S2. Use dynamic and adaptive quantitative modeling methods to generate the initial project scoring matrix and make adaptive adjustments based on historical data and real-time updated data.

[0020] S3. Based on multi-dimensional quantitative modeling, generate a scoring matrix that reflects the characteristics of each dimension of the project.

[0021] Furthermore, the real-time market dynamic monitoring system comprises the following steps:

[0022] S1. First, a multi-source data pool is formed by integrating multiple data sources, including financial market data, industry news, and policy changes. All data are acquired in real time or near real time and integrated into the monitoring system. The real-time data integration is achieved through the function Ω(D, t), which is defined as follows:

[0023]

[0024] Among them, D i is the i-th data source, κ i is the weight of the data source, λ i is the data attenuation factor, φ i (D i ) is the data integration function, T is the monitoring period;

[0025] S2. Use machine learning and big data analysis technology to process and analyze multi-source data in real time, identify potential trends and implicit risks of market changes, and use natural language processing technology to perform semantic analysis on industry news and policy documents to extract key information related to investment projects; market dynamics analysis is achieved through function Γ(M,t), which is defined as follows:

[0026]

[0027] Among them, M j is the j-th type of market data, σ j is the weight coefficient, ψ j (M j ,t) is the time evolution function of market data, η j It is the market impact correction item after semantic analysis.

[0028] Furthermore, the project adaptability analysis module construction includes the following steps:

[0029] S1. Use advanced data fusion technology to deeply integrate real-time market dynamic data with project feature data, where market dynamic data includes stock market fluctuations, industry news, and policy changes, and project feature data includes financial indicators, team background, and market demand forecasts. The data fusion process is implemented through the function Ξ(D, P), which is defined as follows:

[0030]

[0031] Among them, D i is the dynamic data of the ith market, Pj is the feature data of the jth item, ρ ij is the interaction coefficient between market data and project characteristics, and ψ j (P j ) represent the nonlinear influence functions of market data and project characteristic data respectively;

[0032] S2. Use the adaptive learning model to update the performance forecast of the project under different market conditions in real time. The model is implemented by the function Λ(P,t), which is defined as follows:

[0033]

[0034] Among them, γ l is the feature weight, ψ l (P,t) is the time evolution function of the characteristic data, α l and ω l are the adjustment coefficient and the market fluctuation frequency respectively;

[0035] S3. The model is used to dynamically adjust the weights and priorities of project features so that project performance evaluation reflects the current market environment. The project adaptability analysis module also uses a multi-dimensional analysis model to comprehensively consider multiple factors such as market risks, policy impacts, industry trends, and competitive situations, and conducts a multi-dimensional evaluation of the overall performance and future potential of the project under current market conditions.

[0036] Furthermore, the method for ranking investment priorities of projects is as follows:

[0037] S1. First, a comprehensive project scoring matrix is ​​generated, and combined with the project adaptability analysis results to form a multi-dimensional evaluation framework, where the integration of the project scoring matrix and the adaptability analysis results is achieved through a nonlinear mapping function Υ(S, A), which is defined as follows:

[0038]

[0039] Among them, S i represents the i-th dimension of the project rating matrix, A j represents the jth factor of the project adaptability analysis results, ζ ij is the mapping coefficient, ψ i (S) and α j (A j ) are the influence functions of the scoring matrix and the adaptive analysis results, respectively, and κ(t) is the time-related adjustment function;

[0040] S2. Based on the integrated results of S1, use a multi-objective optimization algorithm to sort the investment priorities of the projects.

[0041] Furthermore, the use strategy of the investment priority ranking result includes:

[0042] S1. Generate investment priority ranking results based on investor resource constraints, and dynamically adjust the priority ranking to control the ranking results to meet resource constraints; optimize by introducing constraints and combining investors' personalized preferences including risk tolerance and return targets; the optimization process is implemented by function Ψ(R,C), which is defined as follows:

[0043]

[0044] Among them, R i is the resource constraint, C i is the investment priority ranking result, λ i is the constraint weight coefficient, φ i (R i ) and χ i (C i ) represent the influence functions of resource constraints and sorting results respectively;

[0045] S2. Based on the optimized priority sorting, a resource optimization configuration model is established by combining mixed integer programming (MIP) and dynamic programming (DP), where MIP handles the discreteness of resources and DP is used for dynamic allocation of resources.

[0046] S3. In the resource optimization allocation model, a multi-objective optimization strategy is adopted to optimize multiple objective functions including profit maximization, risk minimization, and liquidity optimization.

[0047] Furthermore, the method for constructing the adjusted investment portfolio includes:

[0048] S1. Generate a large number of predicted market scenarios and project performance paths by introducing Monte Carlo simulation technology, and evaluate the performance of project portfolios under different market conditions by repeatedly simulating scenarios;

[0049] S2. Based on S1, the Bayesian network is introduced to conduct a structured analysis of the uncertainty and potential risks in project investment decisions, capture the dependencies between different projects and market factors, and dynamically adjust the risk assessment results by updating the probability distribution of the network in real time;

[0050] S3. Based on the risk assessment results of the Monte Carlo simulation and Bayesian network, develop a risk-driven portfolio adjustment strategy to identify high-risk projects and recommend reduction or exit, while identifying projects that perform robustly under uncertainty and recommend increasing investment.

[0051] Furthermore, the risk-driven portfolio adjustment strategy construction includes:

[0052] S1. First, we combine Monte Carlo simulation with Bayesian network to conduct multi-model risk assessment, where Monte Carlo simulation is used to generate a large number of market scenarios and analyze the performance of projects under these scenarios. The risk distribution is calculated by the function Quantization, the formula is as follows:

[0053]

[0054] Among them, x i is the market scenario parameter, α i is the risk weight coefficient, Φ i (t) is the performance function that changes with time;

[0055] S2. Secondly, Bayesian networks are used to capture the complex dependencies between projects and market factors, and to achieve structured risk assessment through dynamic updates of probability distributions; the assessment process is implemented through the function The implementation is defined as follows:

[0056]

[0057] in, is the conditional probability distribution related to market risk in the Bayesian network, β j and γ j for risk adjustment parameters;

[0058] S3. Finally, based on the risk assessment results of S1-S2 multi-model, develop a risk-driven portfolio dynamic adjustment strategy, which includes: identifying high-risk projects and using functions It is recommended to reduce holdings or exit, which is defined as follows:

[0059]

[0060] Among them, z is the identification parameter of high-risk projects, δ k To adjust the coefficient; at the same time, identify projects that are robust and resilient under uncertain conditions and recommend increased investment.

[0061] Beneficial effects of the present invention:

[0062] By integrating multiple data sources (such as market dynamics, project characteristics, etc.) and combining machine learning models and big data analysis technology, the present invention can comprehensively analyze all dimensions of a project, thereby generating a scientific and accurate project scoring matrix and priority ranking. This method greatly improves the scientificity and accuracy of investment decisions, enabling investors to better grasp market opportunities.

[0063] Secondly, the introduction of adaptive learning models and multi-objective optimization algorithms can adjust project scores and investment priorities in real time to adapt to the rapidly changing market environment. This dynamic adjustment capability ensures that the investment portfolio can remain in the best state during market fluctuations and helps investors make more informed decisions under uncertain conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 The present invention is a flowchart of the optimization method for project investment decision-making.

[0065] Figure 2 Flowchart of steps for generating a project scoring matrix for the present invention.

[0066] Figure 3 A flowchart of steps for constructing the project adaptability analysis module of the present invention.

[0067] Figure 4 A flow chart of strategic steps is used to prioritize the results of the present invention investments. DETAILED DESCRIPTION

[0068] The specific implementation modes of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0069] As attached Figure 1 As shown, the first step of the present invention includes first collecting and integrating multi-dimensional data related to the project, including the financial data of the project, project background information, market demand forecast, industry trend analysis and project competitiveness assessment. The purpose is to build a comprehensive data foundation covering multiple key areas in the project investment decision-making process. Financial data usually includes the budget, cost structure, expected returns, etc. of the project, which can directly reflect the economic feasibility of the project. Project background information covers the experience, ability, and success rate of previous projects of the members, which have a direct impact on the execution ability of the project. Market demand forecast and industry trend analysis help to understand the potential acceptance and future development space of the project in the target market, while the competitiveness assessment of the project focuses on the unique advantages and disadvantages of the project in the current competitive environment. By collecting and integrating these multi-dimensional data, a more comprehensive and objective basis can be provided for subsequent project analysis and investment decisions, ensuring the scientificity and accuracy of the decision.

[0070] Then, the collected project feature data is deeply analyzed through machine learning models to extract the most critical indicators for project investment decisions. Machine learning models can process large amounts of complex data and identify patterns and relationships hidden behind the data. Specifically, the model will analyze multi-dimensional features including financial data, background, market demand, industry trends, competitiveness, etc. Through the analysis of these features, the model can identify which factors have a significant impact on the success of the project. These important factors are called key indicators. After extracting these key indicators, multi-dimensional quantitative modeling will be carried out for each project. The process of quantitative modeling involves converting the different characteristics of the project into quantifiable values, and comprehensively processing them through mathematical models to generate a project scoring matrix. This scoring matrix can intuitively show the performance of each project in different dimensions. By analyzing these scores, we can further understand the overall strengths and weaknesses of the project.

[0071] The second step is to build a real-time market dynamics monitoring system that continuously monitors market changes by integrating market data from multiple sources and analyzes the potential impact of these changes on the project. First, the system integrates market data from different data sources, including stock market fluctuations, industry news, and policy changes. Stock market fluctuation data provides instant feedback from the financial market, which can reflect investor sentiment and the overall health of the market. Industry news provides important events, trends, and developments in the industry, which can help identify potential opportunities and risks in the industry where the project is located. Policy change data involves adjustments to regulations, laws, and policy orientations issued by the government, which often directly affect the operating environment of the industry and the legitimacy or profitability of the project. By integrating this data, the real-time market dynamics monitoring system can respond quickly to market changes and conduct in-depth analysis. The system analyzes the various potential impacts of market changes on the project, including direct and indirect impacts on project returns, risks, and market demand.

[0072] Next, we developed a project adaptability analysis module that combines market dynamics data with project characteristic data to evaluate the performance and potential of each project under current market conditions. First, we obtained market dynamics data from the real-time market dynamics monitoring system, including stock market fluctuations, industry news, and policy changes. Market dynamics data reflects the changes and trends in the current market environment, which have a direct or indirect impact on the success of the project. Project characteristic data includes information such as the project's financial status, background, market demand forecast, and competitiveness, which reflect the inherent strengths and weaknesses of the project itself. The project adaptability analysis module combines these external market dynamics data with internal project characteristic data, and uses complex algorithms and models to evaluate the performance of each project under current market conditions. This evaluation not only considers the short-term impact of market changes on the project, but also analyzes the potential performance of the project in the long term. Through this comprehensive analysis, the module can identify which projects have high adaptability and success potential under current market conditions, and give corresponding adjustment suggestions.

[0073] The third step is to prioritize the projects by using a multi-objective optimization algorithm based on the project scoring matrix and project adaptability analysis results generated in the previous steps. The project scoring matrix provides performance data for each project in multiple dimensions, including financial indicators, market demand, background, etc., while the project adaptability analysis results reflect the performance potential of the project under current market conditions. The multi-objective optimization algorithm can consider multiple goals at the same time, such as investment return, risk level, project cycle, etc., and perform weighted synthesis on these goals to finally generate a priority ranking list. The ranking list helps investors identify which projects are most worthy of priority investment in the current market environment, thereby optimizing resource allocation. Next, the system updates the investment priorities of these projects in real time based on market changes and project progress. The market is dynamic, so the priority of projects also needs to be adjusted as market conditions change. For example, if a sudden change in market conditions leads to an increase in the risk of certain projects, the system will recalculate and lower the priority of these projects based on the latest data. On the contrary, if a project shows progress beyond expectations during execution, the system will increase its priority to ensure that resources can support these potential projects in a timely manner.

[0074] The fourth step is to consider the resource constraints of investors, which include factors such as capital budget, time, and human resources. Based on these resource constraints and combined with the investment priority ranking results generated previously, the system uses mixed integer programming or dynamic programming algorithms to establish a resource optimization allocation model. Mixed integer programming is an optimization method for dealing with discrete and continuous decision variables, which is suitable for solving problems such as project selection and resource allocation. Dynamic programming decomposes the problem into a series of sub-problems and gradually solves the optimal solution. It is suitable for dealing with dynamic resource allocation problems over time. In this step, the application of these two algorithms enables the resource optimization allocation model to not only meet the resource constraints, but also optimize resource allocation under these conditions. Specifically, the model will screen and calculate in the investment portfolio, and finally select those project portfolios that still have the greatest potential under limited resources for investment.

[0075] The fifth step is to evaluate the uncertainty and potential risks in project investment decisions by introducing Monte Carlo simulation or Bayesian network. Monte Carlo simulation is a method of predicting the distribution of results through a large number of random simulations. It can generate a large number of market scenarios and analyze the performance of projects under these different scenarios, thereby quantifying the risks of projects under various uncertain conditions. Bayesian network is a graph model based on probability theory that can capture the dependencies between projects and market factors and dynamically evaluate the risks of projects by updating the probability distribution in the network. The combination of these two methods can comprehensively evaluate the potential risks of projects from different perspectives and provide investors with more accurate risk predictions. The system dynamically adjusts the investment portfolio based on these risk assessment results. When the risk assessment results show that the risk level of certain projects is too high, the system will recommend reducing or exiting these projects to reduce the risk exposure of the overall investment portfolio. On the contrary, if the assessment results show that certain projects are robust under uncertain conditions and have good risk resistance, the system will recommend increasing investment in these projects to optimize the return potential of the investment portfolio. Through this real-time risk assessment and portfolio adjustment, investors can better manage the uncertainty brought about by market fluctuations, ensure the robustness and flexibility of investment decisions, and thus maximize long-term investment returns in a complex market environment.

[0076] Embodiment 1:

[0077] Combination Figure 2In this embodiment, in the field of new energy, an investment company is considering how to optimize its investment portfolio. The options include multiple new energy projects, such as wind power generation, solar panel production, and electric vehicle battery recycling projects. These projects involve different fields, and have their own characteristics in market prospects, technical risks, and financial performance. In order to make scientific investment decisions, a project scoring matrix is ​​generated to evaluate and sort the investment priorities of these projects.

[0078] First, we collected multi-dimensional feature data about each project, including financial data (such as expected return on investment, operating costs, etc.), technical indicators (such as technology maturity, innovation, etc.), market demand forecasts (such as market growth rate in the next five years), background (such as project experience and professional skills), and policy support (such as government subsidies, tax incentives, etc.). In order to extract the most critical features for investment decisions from these complex data, we introduced a machine learning model and built an adaptive learning mechanism.

[0079] Define the function Ψ(x) to represent the process of extracting key features from a large number of data sets. Set the feature vector of each project to x, which includes multiple dimensions such as financial data and technical indicators. Through adaptive learning, the weight parameters α of different features are obtained. i , these weights reflect the importance of each feature in investment decision-making. Feature extraction function f i (x) represents the extracted project features, such as the project’s innovation score, market growth potential, etc., while g(t) is the time-related feature weight adjustment function, which is used to dynamically adjust the importance of features, taking into account the impact of changes in the market and technological environment on the project within a specific time.

[0080] The function form is as follows:

[0081]

[0082] For example, for a wind power project, its financial data (such as an expected annualized rate of return of 10%) is extracted as one of the key features and assigned a weight of α1 = 0.3; the technical maturity score (the full score is 10 points, and the project scored 8 points) is assigned a weight of α2 = 0.25. Assuming g(t) = 0.05t, the time period is 5 years, that is, T = 5, then:

[0083]

[0084] After calculation, the wind power project's key characteristics score under current market conditions is 5.125 points.

[0085] For a more comprehensive evaluation, a similar calculation is performed for the solar panel production project. For this project, the expected annualized return is set to 12% (weight α1 = 0.35), the technology maturity score is 9 points (weight α2 = 0.2), and the same time adjustment function g(t) = 0.05t is used:

[0086]

[0087] Ψ(x)=4.2+1.8+0.125=6.125

[0088] Through the above calculation, the key characteristics score of the solar panel production project is 6.125 points.

[0089] After calculating and comparing all projects, the system generates a project scoring matrix, showing the performance of each project in multiple dimensions and the overall score. The results show that the overall score of solar panel production projects is higher than that of wind power projects, indicating that under current market conditions, investing in solar projects has higher potential returns and stronger risk resistance.

[0090] The project scoring matrix was used to evaluate the investment priorities of various new energy projects, and a preliminary decision was made to prioritize solar panel production projects. In order to further optimize investment decisions, a dynamic and adaptive quantitative modeling method was used to generate an initial project scoring matrix and make adaptive adjustments based on historical data and real-time updated data.

[0091] A dynamic and adaptive quantitative modeling method is introduced to evaluate and adjust the project scoring matrix in real time. Specifically, the defined project scoring function Φ(y,t) is used, where y represents the project feature set, including financial performance, technical maturity, market demand, etc., and t represents the time variable. The function Φ(y,t) can dynamically adjust the project score in the time dimension, taking into account the impact of market fluctuations, policy changes and other factors on project performance.

[0092] The function form is as follows:

[0093]

[0094] Among them, β j is the characteristic weight coefficient, ranging from 0.1 to 0.5, ω is the market fluctuation frequency, ranging from 0.05 to 0.2, θ j is the phase adjustment parameter of the project feature, ranging from 0 to π. Function h j (y) represents the interactive influence function between characteristics and time, such as the performance of the project’s financial return, market growth potential, etc. over time.

[0095] Taking the solar panel production project as an example, under the current market conditions, the weight coefficient β1 of its financial return (eigenvalue is 12%) is set to 0.35, the weight coefficient β2 of the technology maturity score (eigenvalue is 9 points) is set to 0.2, and the weight coefficient β3 of market demand (eigenvalue is 8 points) is set to 0.25. The market fluctuation frequency ω is set to 0.1, and the phase adjustment parameters θ1, θ2, and θ3 are set to 0.5, 1.0, and 1.5 respectively. The time period is 5 years, that is, the value range of t is 0 to 5 years.

[0096] For financial return characteristics:

[0097]

[0098] For technology maturity characteristics:

[0099]

[0100] Regarding market demand characteristics:

[0101]

[0102] Assume that the calculation is performed at a certain time (such as the third year, t=3 years):

[0103]

[0104] Therefore, the project score of the solar panel production project in the third year is:

[0105] Φ(y,3)=1.003+0.578+0.649=2.230

[0106] Through calculation, the dynamic score of solar panel production projects under different time points and market conditions is obtained. This score not only reflects the current market performance of the project, but also captures the impact of market fluctuations on the project through time adjustment. In the third year, the market environment has changed significantly, and the system adjusts the priority score of the solar project, thereby deciding to increase the capital investment in the project in the next investment plan. On the contrary, if the score of the wind power project decreases during the same period due to a decrease in market demand, the investment in the project will be reduced or the expansion plan of the project will be temporarily shelved.

[0107] We further analyze the performance of multiple projects in different dimensions and generate a more comprehensive project scoring matrix. To this end, we use a multi-dimensional scoring function Θ(z) to comprehensively evaluate the investment value of each project based on multi-dimensional quantitative modeling.

[0108] Use a multi-dimensional scoring function Θ(z), where z is a collection of feature vectors of different dimensions, including multiple key dimensions such as financial data, market demand, and technology maturity. The form of the function Θ(z) is as follows:

[0109]

[0110] Among them, γ k is the dimension weight coefficient, ranging from 0.1 to 0.4; δ k is the correlation adjustment parameter between items, ranging from 0.2 to 0.5; p k (z) is a function that reflects the mutual influence of the characteristics of each dimension and is used to evaluate the comprehensive performance of the project in a specific dimension.

[0111] It is set that three projects need to be scored, including solar panel production projects, wind power generation projects, and electric vehicle battery recycling projects. The following key dimensions are determined: financial data, market demand, and technological maturity. For the solar panel production project, the financial data score is set to 12 points (γ1 = 0.3, δ1 = 0.4), the market demand score is set to 8 points (γ2 = 0.25, δ2 = 0.35), and the technological maturity score is set to 9 points (γ3 = 0.2, δ3 = 0.3).

[0112] Substitute into the scoring function for calculation:

[0113] Θ1(z)=0.3·log(1+e 0.4·12 )

[0114] Θ1(z)≈0.3·log(1+e 4.8 )≈0.3·log(1+121.51)≈0.3·log(122.51)≈0.3·4.81≈1.44Similarly, for the market demand score:

[0115] θ2(z)=0.25·log(1+e 0.35·8 )

[0116] Θ2(z)≈0.25·log(1+e 2.8 )≈0.25·log(1+16.44)≈0.25·log(17.44)≈0.25·2.86

[0117] ≈0.715

[0118] For technology maturity rating:

[0119] Θ3(z)=0.2·log(1+e 0.3·9 )

[0120] Θ3(z)≈0.2·log(1+e 2.7)≈0.2·log(1+14.88)≈0.2·log(15.88)≈0.2·2.76≈0.552

[0121] The total score for the Solar Panel Production project is then:

[0122] Θ(z)=1.44+0.715+0.552=2.707

[0123] Through calculation, the comprehensive scores of the solar panel production project in different dimensions were obtained and compared with other projects. The score of the wind power generation project was set to 2.5 and the score of the electric vehicle battery recycling project was set to 2.3. The results showed that the solar panel production project had the highest score, so it was decided to prioritize the allocation of resources to invest in this project.

[0124] Embodiment 2:

[0125] In this embodiment, a real-time market dynamics monitoring system is introduced to cope with the rapid changes in the market and ensure the timeliness and accuracy of investment decisions. The main function of the system is to form a multi-source data pool by integrating multiple data sources, and to obtain and integrate these data in real time or near real time in order to continuously monitor market dynamics.

[0126] First, multiple data sources are integrated, including financial market data (such as stock prices, interest rates, exchange rates, etc.), industry news (such as technological breakthroughs in the new energy industry, competitor dynamics, etc.), and policy changes (such as government subsidy policies, environmental regulations, etc.). These data sources are extensive and dynamically changing, and can fully reflect the latest changes in the market environment. All of these data are acquired in real time or near real time and integrated into the monitoring system.

[0127] In order to process and integrate these data, a real-time data integration function Ω(D,t) is defined, where D i is the i-th data source, κ i is the weight of the data source (ranging from 0.1 to 0.5), λ i is the data attenuation factor (ranging from 0.01 to 0.1), φ i (D i ) is the data integration function, and T is the monitoring period (set to 1 year). The function Ω(D,t) is in the following form:

[0128]

[0129] It is assumed that three main data sources are being monitored: 1. Stock market index D1, 2. New energy industry news D2, 3. Policy changes D3. The corresponding weight coefficients are κ1=0.4, κ2=0.35, κ3=0.25. The data attenuation factors are λ1=0.05, λ2=0.02, λ3=0.03.

[0130] The integration function φ1(D1) of the stock market index data D1 is set to 5000, the integration function φ2(D2) of the new energy industry news data D2 is set to 300, and the integration function φ3(D3) of the policy change data D3 is set to 200. In this case, the monitoring period T=1 year.

[0131] Substitute these data into the integration function Ω(D,t) for calculation:

[0132]

[0133] First, calculate the three integral terms separately:

[0134]

[0135]

[0136] Finally, the integration result Ω(D,t) is:

[0137] Ω(D,t)≈1902.4+102.92+48.52≈2053.84

[0138] By calculating the real-time data integration function Ω(D,t), we get a comprehensive market dynamics index of 2053.84. This index reflects the overall situation of the current market environment, covering information on the stock market, industry dynamics, policy changes, etc. The system compares this index with the data of the previous cycle and finds that the current market changes have a positive impact on new energy investment. Therefore, it is decided to increase investment in solar panel production projects with stable performance, while continuing to monitor the market dynamics of wind power generation and electric vehicle battery recycling projects.

[0139] Then, in order to more accurately identify the potential trends and implicit risks of market changes and make more effective portfolio adjustments, machine learning and big data analysis technologies are further introduced to process and analyze multi-source data in real time.

[0140] By introducing machine learning models and big data analysis technology, multi-source data that has been integrated into the monitoring system is processed and analyzed in real time. Machine learning models can identify potential trends and hidden risks in market data, while big data analysis technology can process and analyze large-scale historical and real-time data to provide more accurate market forecasts. In addition, natural language processing technology is used to perform semantic analysis on industry news and policy documents to extract key information related to investment projects. These analysis results will be used to dynamically adjust investment strategies to ensure that the investment portfolio can cope with rapid changes in the market.

[0141] Market dynamics analysis is implemented through the function Γ(M,t), where M j is the j-th type of market data, σ j is the weight coefficient (the value range is 0.1 to 0.4), ψ j (M j ,t) is the time evolution function of market data, η j is the market impact correction term after semantic analysis. The function Γ(M,t) is in the following form:

[0142]

[0143] The current monitoring of three major market data is set as follows: 1) stock market data M1, 2) new energy industry news M2, and 3) policy change data M3. These data are processed in real time through big data analysis and machine learning models, and each data source is assigned a different weight coefficient, which is σ1 = 0.35, σ2 = 0.3, and σ3 = 0.25.

[0144] The time evolution function of stock market data ψ1(M1,t) is set to 5000+200t, the time evolution function of industry news data ψ2(M2,t) is set to 300+50t, and the time evolution function of policy change data ψ3(M3,t) is set to 200+30t. The market impact correction terms after semantic analysis are set to η1=100, η2=50, and η3=30 respectively.

[0145] Substitute these data into the market dynamic analysis function Γ(t,t) for calculation, setting the current time point to t = 3 years:

[0146]

[0147] Then, substitute these derivative values ​​and the correction term into Γ(t,t):

[0148]

[0149] By calculating the market dynamics analysis function Γ(M,t), we get the dynamic market analysis index 50. This index reflects the potential risks and trends in the current market environment. Comparing this analysis result with the previous integration result 2053.84, we find that although the overall market trend is relatively positive, there are still certain potential risks. In order to cope with this risk, we decided to maintain investment in solar panel production projects while carefully monitoring market changes in wind power projects so that investment adjustments can be made quickly when necessary.

[0150] Embodiment 3:

[0151] See attached Figure 3 ,In this embodiment, in order to more accurately evaluate the performance and potential of ,each project under the current market conditions, a project adaptability analysis ,module is introduced.,The core of this module is to deeply integrate the real-time market dynamic ,data with the project feature data through advanced data fusion ,technology, so as to achieve a comprehensive evaluation of the ,project.

[0152] Using advanced data fusion technology, real-time market dynamic data (such as stock market fluctuations, industry news, policy changes) and project characteristic data (such as financial indicators, background, market demand forecast) are deeply integrated. The purpose of data fusion is to comprehensively consider the mutual influence of market environment and project characteristics, and provide more accurate project adaptability analysis.

[0153] The data fusion process is implemented by the function Ξ(D,P), and the formula is defined as follows:

[0154]

[0155] Among them, D i is the dynamic data of the ith market, P j is the feature data of the jth item, ρ ij is the interaction coefficient between market data and project characteristics (ranging from 0.1 to 0.5), and ψ j (P j ) represent the nonlinear influence functions of market data and project characteristic data respectively.

[0156] Assume that an investment company is evaluating a solar panel production project. To this end, three types of market dynamics data are collected: 1) stock market fluctuation data D1 (such as stock indexes of related energy sources), 2) industry news data D2 (such as growth expectations for the solar market), and 3) policy change data D3 (such as new government environmental regulations). In addition, three types of project feature data are collected: 1) financial indicators P1 (such as expected return on investment), 2) background P2 (such as technical experience), and 3) market demand forecast P3 (such as the growth rate of solar demand in the next five years).

[0157] Set the nonlinear influence function of stock market volatility data D1 is 1000, the nonlinear influence function of industry news data D2 is 500, the nonlinear impact function of the policy change data D3 is 200. The nonlinear influence function ψ1(P1) of the project's financial indicator P1 is set to 12, the nonlinear influence function ψ2(P2) of the background P2 is set to 9, and the nonlinear influence function ψ3(P3) of the market demand forecast P3 is set to 8. The interaction coefficients between market data and project characteristics are set to ρ 11 =0.4,ρ 12 =0.35,ρ 13 =0.3,ρ 21 =0.25,ρ 22 =0.2,ρ 23 =0.15.

[0158] Calculate the function Ξ(D,P):

[0159]

[0160] Set the second derivatives of the market data to 10, 5, and 2 respectively:

[0161] Ξ(D,P)=(0.4·10·12)+(0.35·5·9)+(0.3·2·8)

[0162] Ξ(D,P)=48+15.75+4.8=68.55

[0163] By calculating the function Ξ(D,P), the comprehensive adaptability score of the solar panel production project under current market conditions was obtained to be 68.55 points. This score combines the interaction between market dynamics and project characteristics and provides a comprehensive evaluation indicator. Based on this score, it was decided to continue to increase investment in the solar project because the project has shown strong adaptability and growth potential in the current market environment.

[0164] Then, we further use the adaptive learning model to update the performance forecast of the project in real time under different market conditions to ensure that investment decisions can flexibly respond to rapid changes in the market. We introduce an adaptive learning model based on the function Λ(P,t), which can dynamically adjust the performance forecast of each project and update it in real time based on the latest market data.

[0165] This model is implemented by the function Λ(P,t), which is defined as follows:

[0166]

[0167] Among them, γ lis the feature weight (ranging from 0.1 to 0.4), ψ l (P,t) is the time evolution function of the characteristic data, α l is the adjustment coefficient (the value range is 0.2 to 0.5), ω l is the market fluctuation frequency (ranging from 0.05 to 0.2). The model aims to dynamically update the performance forecast of the project by adjusting the feature weights in real time and taking into account the impact of market fluctuations.

[0168] The current key monitoring solar panel production project is set, and its key features include financial indicators, market demand and technology maturity. Different weight coefficients are assigned to each feature, namely γ1=0.3 (financial indicators), γ2=0.35 (market demand), and γ3=0.25 (technology maturity). The time evolution function of the characteristic data is set to ψ1(P1,t)=5000+200t(financial indicators), ψ2(P2,t)=300+50t(market demand), ψ3(P3,t)=200+30t(technology maturity), adjustment coefficients α1=0.4, α2=0.3, α3=0.25, and market fluctuation frequencies are ω1=0.1, ω2=0.15, and ω3=0.2 respectively.

[0169] Assume that we need to predict the performance of the project at the current time point t = 3 years. First, calculate the derivative of the time evolution function:

[0170]

[0171] Substitute these data into the adaptive learning model function Λ(P,t):

[0172]

[0173] Calculate the results:

[0174] sin(0.3)≈0.2955, sin(0.45)≈0.4346, sin(0.6)≈0.5646

[0175]

[0176] The calculation of the adaptive learning model function Λ(P,t) yielded a real-time performance forecast of 28.37 for the solar panel production project under current market conditions. This result takes into account the volatility of the market and the evolution of project characteristics over time, providing a dynamically adjusted project performance forecast. Based on this forecast, it was decided to continue to prioritize investments in solar panel production projects while closely monitoring further changes in the market.

[0177] Then, in order to ensure that the project evaluation can accurately reflect the changes in the current market environment, the model is further used to dynamically adjust the project feature weights and priorities. This step aims to comprehensively consider market risks, policy impacts, industry trends, and competitive situations, and conduct a multi-dimensional evaluation of the overall performance and future potential of the project under current market conditions.

[0178] In the project adaptability analysis module, a multi-dimensional analysis model is used, which dynamically evaluates the overall performance of the project through the function Θ(Q). The definition of the function Θ(Q) is as follows:

[0179]

[0180] Among them, Q s is a multi-dimensional vector of project characteristics, covering multiple dimensions such as market risk, policy impact, industry trend, and competitive situation; η s is the dimension weight coefficient (ranging from 0.1 to 0.4), κ s is the characteristic adjustment parameter (the value range is 0.05 to 0.2); T is the evaluation period (set to 1 year).

[0181] A multi-dimensional evaluation of solar panel production projects is set. Four key dimensions are selected: 1) market risk Q1 (such as the risk of increased competition), 2) policy impact Q2 (such as changes in government subsidy policies), 3) industry trend Q3 (such as the growth potential of the solar market), and 4) competitive situation Q4 (such as the development of other alternative energy technologies). For each dimension, the weight coefficients η1=0.3, η2=0.25, η3=0.2, and η4=0.25 are assigned respectively. The feature adjustment parameters are κ1=0.15, κ2=0.1, κ3=0.2, and κ4=0.1 respectively.

[0182] Assuming that the evaluation period is T = 1 year, the market risk characteristic vector Q1 is 6, the policy impact Q2 is 8, the industry trend Q3 is 10, and the competition situation Q4 is 7. Substitute these data into the function Θ(Q) for calculation:

[0183]

[0184] First, the logarithmic terms of each dimension are calculated separately:

[0185] log(1+0.15·6)≈log(1.9)≈0.64

[0186] log(1+0.1·8)≈log(1.8)≈0.59

[0187] log(1+0.2·10)≈log(3)≈1.10

[0188] log(1+0.1·7)≈log(1.7)≈0.53

[0189] Then, substituting these values ​​into Θ(Q):

[0190]

[0191] Θ(Q)=0.692·T=0.692×1=0.692

[0192] Through the calculation of the multidimensional analysis model function Θ(Q), the multidimensional evaluation score of the solar panel production project under the current market conditions was obtained to be 0.692. Based on this score, it was decided to continue to support the solar panel production project, especially because the positive effects of industry trends and policy impacts on the project are more obvious. However, the potential threats of market risks and competitive situation also remind us to remain vigilant and regularly evaluate the changes in these factors so as to adjust the investment strategy in a timely manner when necessary.

[0193] This multi-dimensional analysis can also be applied to other new energy projects, such as wind power projects and electric vehicle battery recycling projects, and it is found that different projects have different performances and prospects in the market environment. By dynamically adjusting feature weights and priorities and combining real-time market data, it is possible to more accurately predict the future performance of projects and flexibly respond to market uncertainties, ultimately achieving investment portfolio optimization and maximum returns.

[0194] Embodiment 4:

[0195] In this embodiment, in the process of continuing to optimize its new energy project investment decision, the investment priority of the projects is ranked to ensure that, under the condition of limited resources, those projects with the best performance under the current market conditions can be invested in priority. To this end, a comprehensive multi-dimensional evaluation framework is adopted, and a multi-objective optimization algorithm is used to generate investment priority ranking.

[0196] Generate a comprehensive project scoring matrix and combine the project suitability analysis results

[0197] First, a comprehensive project scoring matrix is ​​generated, and a multi-dimensional evaluation framework is formed by combining the project adaptability analysis results. This process is achieved through the nonlinear mapping function Υ(S,A), and the formula is defined as follows:

[0198]

[0199] Among them, S i represents the i-th dimension of the project rating matrix, A j represents the jth factor of the project adaptability analysis results, ζ ij is the mapping coefficient (the value range is 0.1 to 0.5), ψ i(S) and α j (A j ) are the scoring matrix and the influence function of the adaptive analysis results, respectively, and κ(t) is the time-related adjustment function.

[0200] The company is evaluating three major new energy projects: solar panel production projects, wind power projects, and electric vehicle battery recycling projects. A scoring matrix for these projects was generated, which included scores for financial performance, market demand, and technology maturity. The results of the project adaptability analysis took into account factors such as market risk, policy support, and industry competitiveness.

[0201] The solar panel production project is set to score 80 points in the financial dimension S1, 75 points in the market demand dimension S2, and 85 points in the technology maturity dimension S3 in the scoring matrix. The adaptability analysis results show that the market risk factor A1 scores 60 points, the policy support factor A2 scores 70 points, and the industry competitiveness factor A3 scores 65 points.

[0202] Substitute these data into the nonlinear mapping function Y(S,A) and set the mapping coefficient ζ 11 =0.3,ζ 12 =0.25, ζ 13 =0.35,ζ 21 =0.4, ζ 22 =0.3, ζ 23 = 0.2. The time-dependent adjustment function κ(t) is assumed to be 0.1 in the period T = 1 year.

[0203] Set the influence function ψ of the scoring matrix i (S) is a simple linear function, then:

[0204]

[0205] Set the influence function α of the adaptive analysis result j (A j )for:

[0206] α1(A1)=1+0.02A1, α2(A2)=1+0.03A2, α3(A3)=1+0.025A3

[0207] Substitute the data into the mapping function Υ(S,A):

[0208]

[0209] Υ(S,A)=0.3·1.2+0.25·1.21+0.35·1.1625+0.1≈0.36+0.3025+0.406875+0.1

[0210] =1.169375

[0211] Using multi-objective optimization algorithms to prioritize projects for investment

[0212] Based on the above integrated Υ(S,A) results, a multi-objective optimization algorithm is used to prioritize the investment of projects. The optimization process is implemented through the objective function Ω(R), which is defined as follows:

[0213]

[0214] Among them, R k is the kth investment return dimension, λ k is the target weight coefficient (ranging from 0.2 to 0.5), φ k (R k ) is the rate of return function, θ k is an adjustment parameter (the value range is 0.05 to 0.2).

[0215] In the analysis, the weight coefficient λ1 assigned to the financial return dimension is 0.4, and the risk-adjusted return rate is R1 = 10%. The return dimension λ2 of market growth potential is 0.35, and the expected return rate is R2 = 15%. The return dimension λ3 of technological innovation is 0.25, and the expected return rate is R3 = 12%. The adjustment parameters θ1 = 0.1, θ2 = 0.15, and θ3 = 0.2.

[0216] Substitute these data into the objective function Ω(R):

[0217]

[0218] Ω(R)=max(0.4·log(101)+0.35·log(101)+0.25·log(61))

[0219] Ω(R)≈max(0.4·4.615+0.35·4.615+0.25·4.110)=max(1.846+1.615+1.0275)

[0220] =max4.4885

[0221] By calculating the nonlinear mapping function Υ(S,A) and the objective function Ω(R), the comprehensive priority score of solar panel production projects in the current market environment was obtained. After comparing the scores of other projects, solar energy projects were rated as the current priority investment projects due to their higher financial return expectations and better policy support.

[0222] At the same time, this analysis result was compared with the priority ranking of other projects, and it was found that although wind power projects have great market potential, they have a relatively low priority due to high uncertainty in policy support. By solving the Pareto optimal solution set, multiple predicted investment priority ranking schemes were generated, and the optimal investment portfolio was finally determined by combining the risk preferences of management and market expectations.

[0223] Embodiment 5:

[0224] Combination Figure 4 ,In this embodiment, a comprehensive investment priority ranking strategy is adopted, and a ,resource optimization allocation model is established by combining the ,mixed integer programming (MIP) and dynamic programming (DP).

[0225] S1. Generate investment priority ranking results based on resource constraints

[0226] First, consider the resource constraints of investors, including capital budget, investment time window, and human resources. After generating the initial investment priority ranking results, the ranking results are dynamically adjusted to meet the resource constraints. To achieve this optimization, constraints are introduced and combined with investors' personalized preferences, including risk tolerance and return targets.

[0227] The optimization process is implemented by the function Ψ(R,C), which is defined as follows:

[0228]

[0229] Among them, R i is the resource constraint, C i is the investment priority ranking result, λ i is the constraint weight coefficient (ranging from 0.1 to 0.5), φ i (R i ) represents the impact function of resource limitation, χ i (C i ) represents the influence function of the sorting result.

[0230] The current resource constraints are set to include capital budget (R1 = 10 million RMB), time limit (R2 = 5 years) and human resource limit (R3 = 20 people). The initial priority ranking result of the solar panel production project is set to C1 = 80, the ranking result of the wind power generation project is set to C2 = 75, and the ranking result of the electric vehicle battery recycling project is set to C3 = 85.

[0231] Set the constraint weight coefficients λ1 = 0.3, λ2 = 0.35, and λ3 = 0.35. Set the resource constraint influence function φ i (R i )for:

[0232] φ1(R1)=0.5R1, φ2(R2)=0.4R2, φ3(R3)=0.3R3

[0233] Set the influence function χ of the sorting result i (C i )for:

[0234] χ1(C1)=1.2C1, χ2(C2)=1.1C2, χ3(C3)=1.15C3

[0235] Substitute the data into the function Ψ(R,C):

[0236]

[0237] S2. Use the combination of MIP and DP to establish a resource optimization configuration model

[0238] Based on the optimized priority ranking, a resource optimization allocation model was established using a combination of mixed integer programming (MIP) and dynamic programming (DP). MIP is used to deal with discrete resource problems, such as whether a project is selected or not, while DP is used to deal with dynamic resource allocation problems, such as the annual investment of funds.

[0239] The joint optimization of the model is achieved through the function Γ(X,t), which is defined as follows:

[0240]

[0241] Among them, X j For the j-th resource allocation decision variable, θ j To optimize the target weight (value range is 0.1 to 0.5), is the dynamic allocation function of resources changing over time, μ j is a discretely selected tuning parameter (ranging from 0.05 to 0.2).

[0242] The resources that need to be allocated within 5 years include funds (X1), time (X2) and manpower (X3). The optimization target weights are set as θ1 = 0.4, θ2 = 0.35, and θ3 = 0.25. The dynamic allocation function of resources that changes over time is set. for:

[0243]

[0244] Set the discrete choice tuning parameter μ j These are 0.1, 0.15, and 0.2 respectively. Substitute these data into the function Γ(X,t) for calculation, setting the capital allocation to 2 million per year, the time allocation to 1 year per year, and the manpower allocation to 5 people per year:

[0245]

[0246] S3. Adopting multi-objective optimization strategy

[0247] In the resource optimization allocation model, a multi-objective optimization strategy is adopted to optimize multiple objective functions including profit maximization, risk minimization, and liquidity optimization. This process is implemented through the function Ω(Z), which is defined as follows:

[0248]

[0249] Among them, Z k For the multi-objective optimization variables of the portfolio, γ k is the target weight coefficient (ranging from 0.2 to 0.5), ρ k To optimize the adjustment parameters (the value range is 0.05 to 0.2).

[0250] The return on investment Z1 for the solar panel production project is set to 10% within three years, the return on the wind power project is set to 12%, and the return on the electric vehicle battery recycling project is set to 9%. The target weight coefficients γ1 = 0.4, γ2 = 0.35, γ3 = 0.25 are set, and the optimization adjustment parameters ρ1 = 0.15, ρ2 = 0.1, ρ3 = 0.2.

[0251] Substituting these data into the function Ω(Z):

[0252] Ω(Z)=max(0.4·log(1+0.15·10)+0.35·log(1+0.1·12)+0.25·log(1+0.2·9))

[0253] Ω(Z)≈max(0.4·log2.5+0.35·log2.2+0.25·log2.8)

[0254] =max(0.4·0.9163+0.35·0.7885+0.25·1.0296)=max 1.7344

[0255] Through the above three-step optimization process, the results of the investment portfolio decision under the resource optimization allocation and multi-objective optimization strategy were obtained. The solar panel production project ranked first in the optimized investment priority ranking, followed by the wind power generation project, and the electric vehicle battery recycling project ranked third. This decision took into account the current resource constraints, the financial returns of the project, the market risks, and the future market potential, and combined mixed integer programming and dynamic programming to ensure that resources can be flexibly allocated in a complex market environment, maximizing investment returns and minimizing risks.

[0256] Embodiment 6:

[0257] In this embodiment, a risk-driven portfolio adjustment strategy is introduced to cope with market uncertainty and ensure the stability and profitability of the portfolio. A dynamic portfolio adjustment strategy is developed by combining Monte Carlo simulation and Bayesian network multi-model risk assessment methods.

[0258] S1. Multi-model risk assessment via Monte Carlo simulation

[0259] First, a large number of market scenarios are generated through Monte Carlo simulation, and the performance of each project under these scenarios is analyzed. The purpose of this simulation process is to quantify the risk distribution of the project under different market conditions. The risk distribution is calculated by the function To implement, the formula is defined as follows:

[0260]

[0261] Among them, x i is the market scenario parameter, α i is the risk weight coefficient (ranging from 0.1 to 0.4), Φ i (t) is the performance function that changes with time.

[0262] It is assumed that three major projects are being evaluated: solar panel production project, wind power generation project and electric vehicle battery recycling project. Through Monte Carlo simulation, parameter data under different market scenarios are generated, such as economic growth rate, energy price fluctuations, etc. In the 1000 market scenarios, the market scenario parameter x1 of the solar panel production project is 0.03, the market scenario parameter x2 of the wind power generation project is 0.02, and the market scenario parameter x3 of the electric vehicle battery recycling project is 0.01.

[0263] Assume risk weight coefficients α1 = 0.3, α2 = 0.2, α3 = 0.15, and time-dependent performance function Φ i (t) is:

[0264] Φ1(t)=10·t, Φ2(t)=8·log(t+1), Φ3(t)=5·t 2

[0265] Substitute data into the function middle:

[0266]

[0267] S2. Structured risk assessment via Bayesian networks

[0268] Based on the generated risk distribution, Bayesian network is used to capture the complex dependencies between projects and market factors, and structured risk assessment is achieved through dynamic updating of probability distribution. To implement, the formula is defined as follows:

[0269]

[0270] in, is the conditional probability distribution related to market risk in the Bayesian network, β j and γ j are risk adjustment parameters, with values ​​ranging from 0.2 to 0.5 and 0.1 to 0.3, respectively.

[0271] The risk adjustment parameters used in the analysis are set to β1 = 0.4, β2 = 0.35, β3 = 0.25, and the conditional probability distribution function Based on previous risk assessment As a result, adjustments are made and γ1=0.2, γ2=0.25, and γ3=0.15 are set.

[0272] Substitute data into the function middle:

[0273]

[0274] S3. Develop risk-driven portfolio dynamic adjustment strategies

[0275] Based on the above multi-model risk assessment results, a risk-driven portfolio dynamic adjustment strategy is developed. This strategy includes identifying high-risk projects and using functions It is recommended to reduce holdings or exit, and at the same time identify projects with stable performance and risk resistance and recommend increasing investment. The formula is defined as follows:

[0276]

[0277] Among them, z is the identification parameter of high-risk projects, δ k is the adjustment coefficient (the value range is 0.1 to 0.3).

[0278] In the analysis, it is decided to set the high risk identification parameter z1 of the solar panel production project to 10% and the adjustment coefficient δ1 to 0.2. By solving and integrating the derivative of , we can conclude that the investment proportion in solar panel production projects should be reduced, while the investment in wind power generation projects should be increased.

[0279] Through Monte Carlo simulation and Bayesian network multi-model risk assessment, the risk distribution of each project in different market scenarios was obtained. Solar panel production projects show higher risks in certain market scenarios, especially in scenarios where policy changes lead to market fluctuations. As a result, it was decided to gradually reduce investment in solar panel production projects and reallocate resources to wind power projects, which showed greater stability under uncertain conditions.

[0280] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. The optimization method of project investment decision-making is characterized by The following steps are involved: S1. Collect and integrate multi-dimensional data related to the project, including financial data of the project, background information of the project team, market demand forecast, industry trend analysis and competitiveness assessment of the project; S1.

1. Analyze project characteristics through machine learning models, extract key indicators, and conduct multi-dimensional quantitative modeling for each project to generate a project scoring matrix; S2. Build a real-time market dynamics monitoring system to analyze the potential impact of market changes on projects by integrating market data, including stock market fluctuations, industry news and policy changes; S2.

1. Develop a project adaptability analysis module that combines market dynamics data with project characteristic data to assess the performance and potential of each project under current market conditions; S3. Based on the project scoring matrix and the project adaptability analysis results, the investment priorities of the projects are sorted using a multi-objective optimization algorithm; S3.

1. Update the investment priorities of the projects in real time according to market changes and project progress; S4. Based on the resource constraints of investors, the investment priority ranking results are used to establish a resource optimization allocation model using mixed integer programming or dynamic programming algorithms. Under the condition of limited resources, the resource allocation strategy is optimized and a potential project portfolio is selected for investment; S5. Introduce Monte Carlo simulation or Bayesian network to evaluate the uncertainty and potential risks in the investment decision of the project; S5.

1. Adjust the investment portfolio according to the risk assessment results.

2. The method for optimizing project investment decisions according to claim 1, characterized in that Generating the project scoring matrix comprises the following steps: S1. Use machine learning models to extract and select project features, and build an adaptive learning mechanism; define the function Ψ(x) to represent the process of extracting key features from a large number of data sets, where x is the feature vector of the project, and the form of Ψ(x) is as follows: Among them, α i is the weight parameter obtained by adaptive learning, f i (x) represents the feature extraction function, g(t) is the time-dependent feature weight adjustment function; S2. Use dynamic and adaptive quantitative modeling methods to generate the initial project scoring matrix and make adaptive adjustments based on historical data and real-time updated data. S3. Based on multi-dimensional quantitative modeling, generate a scoring matrix that reflects the characteristics of each dimension of the project.

3. The method for optimizing project investment decisions according to claim 1, characterized in that The real-time market dynamic monitoring system comprises the following steps: S1. First, a multi-source data pool is formed by integrating multiple data sources, including financial market data, industry news, and policy changes. All data are acquired in real time or near real time and integrated into the monitoring system. The real-time data integration is achieved through the function Ω(D, t), which is defined as follows: Among them, D i is the i-th data source, κ i is the weight of the data source, λ i is the data attenuation factor, φ i (D i ) is the data integration function, T is the monitoring period; S2. Use machine learning and big data analysis technology to process and analyze multi-source data in real time, identify potential trends and implicit risks of market changes, and use natural language processing technology to perform semantic analysis on industry news and policy documents to extract key information related to investment projects; market dynamics analysis is achieved through function Γ(M,t), which is defined as follows: Among them, M j is the j-th type of market data, σ j is the weight coefficient, ψ j (M j ,t) is the time evolution function of market data, η j It is the market impact correction item after semantic analysis.

4. The method for optimizing project investment decisions according to claim 1, characterized in that The project adaptability analysis module construction includes the following steps: S1. Use advanced data fusion technology to deeply integrate real-time market dynamic data with project feature data, where market dynamic data includes stock market fluctuations, industry news, and policy changes, and project feature data includes financial indicators, team background, and market demand forecasts. The data fusion process is implemented through the function Ξ(D, P), which is defined as follows: Among them, D i is the dynamic data of the ith market, P j is the feature data of the jth item, ρ ij is the interaction coefficient between market data and project characteristics, and ψ j (P j ) represent the nonlinear influence functions of market data and project characteristic data respectively; S2. Use the adaptive learning model to update the performance forecast of the project under different market conditions in real time. The model is implemented by the function Λ(P,t), which is defined as follows: Among them, γ l is the feature weight, ψ l (P,t) is the time evolution function of the characteristic data, α l and ω l are the adjustment coefficient and the market fluctuation frequency respectively; S3. The model is used to dynamically adjust the weights and priorities of project features so that project performance evaluation reflects the current market environment. The project adaptability analysis module also uses a multi-dimensional analysis model to comprehensively consider multiple factors such as market risks, policy impacts, industry trends, and competitive situations, and conducts a multi-dimensional evaluation of the overall performance and future potential of the project under current market conditions.

5. The method for optimizing project investment decisions according to claim 1, characterized in that The method for prioritizing projects for investment is as follows: S1. First, a comprehensive project scoring matrix is ​​generated, and combined with the project adaptability analysis results to form a multi-dimensional evaluation framework, where the integration of the project scoring matrix and the adaptability analysis results is achieved through a nonlinear mapping function Υ(S, A), which is defined as follows: Among them, S i represents the i-th dimension of the project rating matrix, A j represents the jth factor of the project adaptability analysis results, ζ ij is the mapping coefficient, ψ i (S) and α j (A j ) are the influence functions of the scoring matrix and the adaptive analysis results, respectively, and κ(t) is the time-related adjustment function; S2. Based on the integrated results of S1, use a multi-objective optimization algorithm to sort the investment priorities of the projects.

6. The method for optimizing project investment decisions according to claim 1, characterized in that Strategies for using the investment prioritization results include: S1. Generate investment priority ranking results based on investor resource constraints, and dynamically adjust the priority ranking to control the ranking results to meet resource constraints; optimize by introducing constraints and combining investors' personalized preferences including risk tolerance and return targets; the optimization process is implemented by function Ψ(R,C), which is defined as follows: Among them, R i is the resource constraint, C i is the investment priority ranking result, λ i is the constraint weight coefficient, φ i (R i ) and χ i (C i ) represent the influence functions of resource constraints and sorting results respectively; S2. Based on the optimized priority sorting, a resource optimization configuration model is established by combining mixed integer programming (MIP) and dynamic programming (DP), where MIP handles the discreteness of resources and DP is used for dynamic allocation of resources. S3. In the resource optimization allocation model, a multi-objective optimization strategy is adopted to optimize multiple objective functions including profit maximization, risk minimization, and liquidity optimization.

7. The method for optimizing project investment decisions according to claim 1, characterized in that The method for constructing the adjustment investment portfolio includes: S1. Generate a large number of predicted market scenarios and project performance paths by introducing Monte Carlo simulation technology, and evaluate the performance of project portfolios under different market conditions by repeatedly simulating scenarios; S2. Based on S1, the Bayesian network is introduced to conduct a structured analysis of the uncertainty and potential risks in project investment decisions, capture the dependencies between different projects and market factors, and dynamically adjust the risk assessment results by updating the probability distribution of the network in real time; S3. Based on the risk assessment results of the Monte Carlo simulation and Bayesian network, develop a risk-driven portfolio adjustment strategy to identify high-risk projects and recommend reduction or exit, while identifying projects that perform robustly under uncertainty and recommend increasing investment.

8. The method for optimizing project investment decisions according to claim 7, characterized in that The risk-driven portfolio adjustment strategy construction includes: S1. First, we combine Monte Carlo simulation with Bayesian network to conduct multi-model risk assessment, where Monte Carlo simulation is used to generate a large number of market scenarios and analyze the performance of projects under these scenarios. The risk distribution is calculated by the function Quantization, the formula is as follows: Among them, x i is the market scenario parameter, α i is the risk weight coefficient, Φ i (t) is the performance function that changes with time; S2. Secondly, Bayesian networks are used to capture the complex dependencies between projects and market factors, and to achieve structured risk assessment through dynamic updates of probability distributions; the assessment process is implemented through the function The implementation is defined as follows: in, is the conditional probability distribution related to market risk in the Bayesian network, β j and γ j for risk adjustment parameters; S3. Finally, based on the risk assessment results of S1-S2 multi-model, develop a risk-driven portfolio dynamic adjustment strategy, which includes: identifying high-risk projects and using functions It is recommended to reduce holdings or exit, which is defined as follows: Among them, z is the identification parameter of high-risk projects, δ k To adjust the coefficient; at the same time, identify projects that are robust and resilient under uncertain conditions and recommend increased investment.

Citation Information

Cited By

  • CPS-based engineering project full-period management method and system

    CN119398464A

  • Novel scientific research project intelligent management and control platform

    CN119919084A

  • E-commerce promotion method based on enterprise condition and market demand design

    CN120258873A

  • An e-commerce promotion method based on enterprise conditions and market demand design

    CN120258873B

  • Evaluation model system and method for upgrading and reconstruction value of parking lot

    CN120373955A