Asset use analysis method and system based on intelligent prediction and optimization algorithm

Abstract

The invention discloses an asset use analysis method and system based on an intelligent prediction and optimization algorithm, and aims to solve the problems of low efficiency, insufficient prediction capability and unscientific purchase decision in existing asset management. According to the invention, asset use records, maintenance information and purchase history are obtained through the data acquisition module; constructing an asset use prediction model by using time sequence prediction, regression analysis and machine learning algorithms, and predicting future asset demand quantity and use trend; generating an optimized purchase plan and an asset allocation scheme in combination with an inventory state, an asset life cycle and a budget constraint; and the prediction result and the optimization suggestion are displayed through a visual tool, and support is provided for decision making. According to the invention, efficient planning and purchase optimization of asset use are realized, and the method can be widely applied to the field of enterprise asset management and supply chain optimization.

Description

Technical Field

[0001] The present invention belongs to the technical field of time series prediction, regression analysis and machine learning, and in particular relates to an asset usage analysis method and system based on intelligent prediction and optimization algorithms. Background Art

[0002] In modern asset management, asset utilization efficiency and scientific procurement decisions play a vital role in optimizing enterprise operating costs and resource allocation. However, traditional asset management methods rely on manual experience and static data analysis, which has many limitations:

[0003] (1) Difficulty in forecasting asset demand: Traditional forecasting methods often lack in-depth analysis of historical data and cannot effectively capture the cyclical, trend-based, and sudden changes in asset usage, resulting in irrational resource allocation and asset oversupply or shortage problems;

[0004] (2) Purchasing decisions rely on experience: Current purchasing plans are usually based on fixed cycles and experience, ignoring actual demand changes and inventory status, which can easily lead to over-purchasing or under-supply, increasing the company's operating costs;

[0005] (3) Lack of intelligent management tools: Most existing systems are limited to asset records and simple statistical analysis. They lack the ability to conduct comprehensive analysis and dynamic prediction of usage data, inventory data, and procurement information, and are unable to provide effective decision-making support for enterprises.

[0006] (4) Low data utilization: Enterprises have accumulated a large amount of asset usage records, inventory data, and financial information. However, due to the dispersion of data and non-uniform formats, it is difficult to conduct efficient integration and in-depth analysis, and the potential value of the data has not been fully tapped. Summary of the invention

[0007] In response to the above-mentioned problems, the present invention proposes an asset usage analysis method and system based on intelligent prediction and optimization algorithms, which aims to achieve accurate prediction of asset usage demand and scientific optimization of procurement plans by combining time series prediction, regression analysis and optimization algorithms, and can dynamically integrate asset usage records, inventory status and financial information, and provide managers with intuitive decision support through visualization tools, so as to help enterprises reduce operating costs, improve asset utilization efficiency, optimize resource allocation, and promote the development of asset management towards intelligent and data-driven directions.

[0008] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is an asset usage analysis method and system based on intelligent prediction and optimization algorithm, which includes the following steps:

[0009] S1. Obtain asset usage records and inventory information from the asset management system;

[0010] S2. Obtain purchase information and depreciation information from the financial system;

[0011] S3. Based on the asset usage records obtained in step S1 and the procurement information obtained in step S2, use the time series model and regression analysis method to predict the future usage and demand trend of the asset;

[0012] S4. Based on the prediction result of step S3, the inventory information obtained in step S1 and the depreciation information obtained in step S2, a procurement and configuration plan is generated in combination with an optimization algorithm;

[0013] S5. Use visualization tools to display the prediction and optimization results of steps S3 and S4 to provide support for asset management and procurement decisions.

[0014] Furthermore, the specific method of obtaining asset usage records and inventory information in step S1 is:

[0015] S11. Connect to the asset management system and automatically obtain asset usage records (including usage frequency, usage duration, failure rate, etc.) data set S1 and real-time inventory information (current inventory, assets in transit, usage status) data set S2 through API or database query;

[0016] S12. Based on the asset usage record dataset S1 and the real-time inventory information dataset S2, remove missing values ​​and unreasonable values, unify the format (such as date, amount and other fields), associate various asset records through primary keys (such as asset numbers), ensure data consistency, and form the asset usage dataset S3;

[0017] Furthermore, the specific method of obtaining the purchase information and depreciation information in step S2 is:

[0018] S21. Import the historical purchase order data set S4 from the financial system, including the purchase date, purchase quantity, supplier information and purchase amount, and obtain the depreciation policy (such as straight-line method, double declining balance method) and specific depreciation details (such as net asset value, accumulated depreciation) data set S5;

[0019] S22. Calculate the residual value of the asset according to the depreciation algorithm to form the residual value data set S6 of the asset. Use the asset number or name to associate the historical purchase order data set S4, the asset usage data set S3 and the residual value data set S6 of the asset to form the asset operation information data set S7;

[0020] Furthermore, the specific method of using the time series model and regression analysis method to predict the future usage and demand trend of assets in step S3 is:

[0021] S31. Use the historical asset usage data set S3 to establish a time series difference autoregressive moving average ARIMA model for short-term forecasting. The three key parameters of the ARIMA model are: (p, d, q), where p (autoregressive order): depends on the data value of the previous p time points; d (number of differences): the number of difference operations required for data stabilization; q (moving average order): depends on the error value of the previous q time points. Use ADF (Augmented Dickey-Fuller) or KPSS test to determine whether the time series is stable. If it is not stable, a difference operation is required:

[0022] Y′ t =Y t -Y t-1

[0023] Then, draw the ACF (autocorrelation function, used to identify the moving average order q) and PACF (partial autocorrelation function, used to identify the autoregressive order p), and determine the p and q values ​​according to the cutoff point. Use actual data to compare the model prediction results and calculate the evaluation indicators:

[0024]

[0025] (Mean Square Error (MSE))

[0026]

[0027] (Mean Absolute Error (MAE))

[0028] Based on the forecast results, determine the future asset demand and usage frequency. The ARIMA model achieves accurate forecasting of asset usage through the stabilization of historical data, autocorrelation analysis and parameter optimization. Combined with residual analysis and performance evaluation, the reliability of the model and the availability of forecast results can be ensured, providing a scientific basis for asset usage planning and procurement optimization.

[0029] S32. Use the historical asset usage dataset S3 to build a time series long short-term memory network LSTM model for short-term prediction. First, the data needs to be normalized. In order to accelerate model convergence, the asset usage is normalized:

[0030]

[0031] Convert the time series into the input format of the LSTM model (sequence-prediction problem): input-historical data of fixed window length (such as usage in the past 30 days), output-predicted future usage (such as the 31st day). Divide the data into a training set (for model training (about 70% to 80%)), a validation set (to adjust model parameters (about 10% to 15%)), and a test set (to evaluate model performance (about 10% to 15%)). Use the training data to fit the model, and use the validation data to monitor overfitting. The monitoring indicator loss function is:

[0032]

[0033] By processing historical asset usage records through the LSTM model, we can effectively capture long-term dependencies and complex patterns in time series and generate high-precision future usage forecasts. This provides a scientific basis for asset usage planning and procurement optimization, and improves resource utilization efficiency.

[0034] S33. Use the historical asset usage dataset S3 to build a forecasting model to quantify the impact of various factors on asset demand and provide support for future asset usage planning.

[0035] First, the data needs to be preprocessed. Interpolation or mean filling is used for missing data, the Z-score method is used to detect and remove outliers, and numerical data is standardized to avoid the impact of feature magnitude differences on the model:

[0036]

[0037] Where μ is the mean and σ is the standard deviation. Determine the predictor variables (X): time factors (such as weeks, months); asset categories (such as equipment, tools); frequency of use, inventory status. Determine the target variable (Y): future usage of the asset (such as demand in the next week or the next quarter).

[0038] Then, different regression analysis methods are used according to different application scenario requirements. When there is a linear relationship between asset demand and influencing factors, such as demand gradually increasing over time, linear regression is suitable:

[0039] Y=β0+β1X1+β2X2+…+β n X n +∈

[0040] Where, Y: predicted asset usage; X: i : Influencing factors (such as time, inventory status); β i : regression coefficient, quantifying the impact of each feature on asset demand; ∈: error term.

[0041] When asset demand is affected by multiple factors (such as seasonal fluctuations and inventory levels), it is more effective to use multiple linear regression:

[0042]

[0043] When the relationship between asset demand and some factors is nonlinear (such as seasonal fluctuations), polynomial regression is applicable:

[0044]

[0045] In high-dimensional data (such as those containing multiple asset types and external factors), select key features, use ridge regression and LASSO regression, and introduce regularization terms to solve multicollinearity or overfitting problems:

[0046] Ridge regression: Minimize||Y-Xβ|| 2 +λ||β|| 2

[0047] LASSO: Minimize||Y-Xβ|| 2 +λ||β||1

[0048] Furthermore, the specific method of generating the procurement and configuration plan by the optimization algorithm in step S4 is:

[0049] Input variables: forecasted asset usage demand, current inventory status, and asset depreciation;

[0050] Optimization goal: Minimize total cost (purchase cost + maintenance cost + depreciation loss) to meet demand and inventory security constraints;

[0051] Generate a procurement plan: determine procurement time, procurement quantity and preferred suppliers;

[0052] Optimize asset allocation: Dynamically adjust the allocation of assets among departments to reduce idleness or shortages.

[0053] Linear Programming Model:

[0054] Minimize: C total =C purchase +C maintenance +C depreciation

[0055] Furthermore, the prediction and optimization results are displayed by visualization tools in step S5 to provide support for decision making in the following specific ways:

[0056] Data visualization: Use charts (such as line charts, bar charts, and heat maps) to display predicted asset usage and demand trends, and display optimized procurement plans and asset allocation plans (such as optimal procurement time and quantity);

[0057] Dynamic interaction: Provides a user interaction interface, supports conditional filtering (such as time period, asset type), and supports data export (such as generating procurement reports) BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flowchart of the steps of an asset usage analysis method based on intelligent prediction and optimization algorithms;

[0059] Figure 2 This is a structural diagram of an asset usage analysis system based on intelligent prediction and optimization algorithms. DETAILED DESCRIPTION

[0060] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0061] like Figure 1 As shown, the present invention discloses an asset usage analysis method based on intelligent prediction and optimization algorithm, the method comprising the following steps:

[0062] S1. Obtain asset usage records and inventory information from the asset management system;

[0063] S2. Obtain purchase information and depreciation information from the financial system;

[0064] S3. Based on the asset usage records obtained in step S1 and the procurement information obtained in step S2, use the time series model and regression analysis method to predict the future usage and demand trend of the asset;

[0065] S4. Based on the prediction result of step S3, the inventory information obtained in step S1 and the depreciation information obtained in step S2, a procurement and configuration plan is generated in combination with an optimization algorithm;

[0066] S5. Use visualization tools to display the prediction and optimization results of steps S3 and S4 to provide support for asset management and procurement decisions.

[0067] The specific method of obtaining asset usage records and inventory information in step S1 is:

[0068] S11. Connect to the asset management system and automatically obtain asset usage records (including usage frequency, usage duration, failure rate, etc.) data set S1 and real-time inventory information (current inventory, assets in transit, usage status) data set S2 through API or database query;

[0069] S12. Based on the asset usage record dataset S1 and the real-time inventory information dataset S2, remove missing values ​​and unreasonable values, unify the format (such as date, amount and other fields), associate various asset records through primary keys (such as asset numbers), ensure data consistency, and form the asset usage dataset S3;

[0070] The specific method of obtaining the purchase information and depreciation information in step S2 is:

[0071] S21. Import the historical purchase order data set S4 from the financial system, including the purchase date, purchase quantity, supplier information and purchase amount, and obtain the depreciation policy (such as straight-line method, double declining balance method) and specific depreciation details (such as net asset value, accumulated depreciation) data set S5;

[0072] S22. Calculate the residual value of the asset according to the depreciation algorithm to form the residual value data set S6 of the asset. Use the asset number or name to associate the historical purchase order data set S4, the asset usage data set S3 and the residual value data set S6 of the asset to form the asset operation information data set S7;

[0073] The specific method of using the time series model and regression analysis method to predict the future usage and demand trend of assets in step S3 is:

[0074] S31. Using the historical asset usage data set S3, establish a time series difference autoregressive moving average ARIMA model for short-term forecasting. The three key parameters of the ARIMA model are: (p, d, q), where p (autoregressive order): depends on the previous p The data value at each time point; d (number of differences): the number of difference operations required to stabilize the data; q (moving average order): the error value that depends on the previous q time points. Use ADF (Augmented Dickey-Fuller) or KPSS test to determine whether the time series is stable. If it is not stable, a difference operation is required:

[0075] Y′ t =Y t -Y t-1

[0076] Then, draw the ACF (autocorrelation function, used to identify the moving average order q) and PACF (partial autocorrelation function, used to identify the autoregressive order p), and determine the p and q values ​​according to the cutoff point. Use actual data to compare the model prediction results and calculate the evaluation indicators:

[0077]

[0078] (Mean Square Error (MSE))

[0079]

[0080] (Mean Absolute Error (MAE))

[0081] Based on the forecast results, determine the future asset demand and usage frequency. The ARIMA model achieves accurate forecasting of asset usage through the stabilization of historical data, autocorrelation analysis and parameter optimization. Combined with residual analysis and performance evaluation, the reliability of the model and the availability of forecast results can be ensured, providing a scientific basis for asset usage planning and procurement optimization.

[0082] S32. Use the historical asset usage dataset S3 to build a time series long short-term memory network LSTM model for short-term prediction. First, the data needs to be normalized. In order to accelerate model convergence, the asset usage is normalized:

[0083]

[0084] Convert the time series into the input format of the LSTM model (sequence-prediction problem): input-historical data of fixed window length (such as usage in the past 30 days), output-predicted future usage (such as the 31st day). Divide the data into a training set (for model training (about 70% to 80%)), a validation set (to adjust model parameters (about 10% to 15%)), and a test set (to evaluate model performance (about 10% to 15%)). Use the training data to fit the model, and use the validation data to monitor overfitting. The monitoring indicator loss function is:

[0085]

[0086] By processing historical asset usage records through the LSTM model, we can effectively capture long-term dependencies and complex patterns in time series and generate high-precision future usage forecasts. This provides a scientific basis for asset usage planning and procurement optimization, and improves resource utilization efficiency.

[0087] S33. Use the historical asset usage dataset S3 to build a forecasting model to quantify the impact of various factors on asset demand and provide support for future asset usage planning.

[0088] First, the data needs to be preprocessed. Interpolation or mean filling is used for missing data, the Z-score method is used to detect and remove outliers, and numerical data is standardized to avoid the impact of feature magnitude differences on the model:

[0089]

[0090] Where μ is the mean and σ is the standard deviation. Determine the predictor variables (X): time factors (such as weeks, months); asset categories (such as equipment, tools); frequency of use, inventory status. Determine the target variable (Y): future usage of the asset (such as demand in the next week or the next quarter).

[0091] Then, different regression analysis methods are used according to different application scenario requirements. When there is a linear relationship between asset demand and influencing factors, such as demand gradually increasing over time, linear regression is suitable:

[0092] Y=β0+β1X1+β2X2+…+β n X n +∈

[0093] Where, Y: predicted asset usage; X: i : Influencing factors (such as time, inventory status); β i : regression coefficient, quantifying the impact of each feature on asset demand; ∈: error term.

[0094] When asset demand is affected by multiple factors (such as seasonal fluctuations and inventory levels), it is more effective to use multiple linear regression:

[0095]

[0096] When the relationship between asset demand and some factors is nonlinear (such as seasonal fluctuations), polynomial regression is applicable:

[0097]

[0098] In high-dimensional data (such as those containing multiple asset types and external factors), select key features, use ridge regression and LASSO regression, and introduce regularization terms to solve multicollinearity or overfitting problems:

[0099] Ridge regression: Minimize||Y-Xβ|| 2 +λ||β|| 2

[0100] LASSO: Minimize||Y-Xβ|| 2 +λ||β||1

[0101] The specific method of generating the procurement and configuration plan by the optimization algorithm in step S4 is:

[0102] Input variables: forecasted asset usage demand, current inventory status, and asset depreciation;

[0103] Optimization goal: Minimize total cost (purchase cost + maintenance cost + depreciation loss) to meet demand and inventory security constraints;

[0104] Generate a procurement plan: determine procurement time, procurement quantity and preferred suppliers;

[0105] Optimize asset allocation: Dynamically adjust the allocation of assets among departments to reduce idleness or shortages.

[0106] Linear Programming Model:

[0107] Minimize: C total =C purchase +C maintenance +C depreciation

[0108] The specific method of displaying the prediction and optimization results by using visualization tools in step S5 to support decision making is as follows:

[0109] Data visualization: Use charts (such as line charts, bar charts, and heat maps) to display predicted asset usage and demand trends, and display optimized procurement plans and asset allocation plans (such as optimal procurement time and quantity);

[0110] Dynamic interaction: Provides a user interaction interface, supports conditional filtering (such as time period, asset type), and supports data export (such as generating procurement reports)

[0111] In order to realize the above-mentioned asset usage analysis method based on intelligent prediction and optimization algorithm, the present invention also includes an asset usage analysis system based on intelligent prediction and optimization algorithm.

[0112] like Figure 2 As shown, the present invention also discloses an asset usage analysis system based on intelligent prediction and optimization algorithm, and the specific functional modules include:

[0113] A data collection module, which is used to automatically collect asset-related information data from different data sources;

[0114] Data modeling module, which builds prediction and optimization models based on collected historical data;

[0115] Predictive analysis module, used to predict future asset usage and demand trends;

[0116] The optimization planning module generates optimized asset procurement and configuration plans based on forecast results, inventory status and budget constraints;

[0117] Visual display module, which displays prediction and optimization results through a graphical interface;

[0118] Data storage and management module, which uniformly stores asset usage records, prediction models and optimization plans to ensure data security and traceability;

[0119] The system management module provides background management functions for system operation, including user management, permission setting and log recording.

[0120] The data acquisition module includes:

[0121] Obtain asset usage records and inventory information from the asset management system; obtain procurement history, budget data and depreciation information from the financial system; obtain external factors that affect asset demand (such as seasonal data, business activity plans, etc.) from external data sources. Support multiple data interfaces (API, database connection, etc.), and realize data cleaning, format standardization and real-time synchronization to ensure data integrity and consistency.

[0122] The data modeling module includes:

[0123] Use time series forecasting models to capture trends and cyclical changes in asset usage; use regression analysis models to quantify key factors affecting asset demand; use optimization algorithm models to comprehensively consider inventory, budget and demand to generate procurement and allocation plans. Flexible selection of algorithm models (such as ARIMA, LSTM, linear regression, genetic algorithm) supports dynamic model updates, and real-time adjustment of forecasts and optimization results based on new data.

[0124] The predictive analysis module includes:

[0125] Use short-term forecasts to meet the forecast of immediate demand changes; use long-term trend analysis to support strategic resource planning. Support comprehensive forecasts based on multi-dimensional data, taking into account the dynamic changes of historical usage records, inventory status and business plans. The forecast results are output in the form of numerical values ​​and charts for intuitive understanding and use.

[0126] The optimization planning module includes:

[0127] Generate optimized asset procurement and configuration plans based on forecast results, inventory status and budget constraints, including procurement optimization: determine the best procurement time, procurement quantity and preferred suppliers; resource allocation optimization: dynamically adjust the allocation of assets between departments to reduce idle or shortage situations. Use optimization methods such as linear programming and genetic algorithms to meet the global optimal solution under multiple constraints. Support flexible adjustment of optimization goals (such as minimizing costs or maximizing resource utilization).

[0128] The visualization module includes:

[0129] Display asset usage trend charts and demand change curves; present procurement plans and resource allocation plans in tables or graphics; support users to filter data by time range, asset category and other dimensions according to needs. Provide multiple data visualization methods (such as line charts, bar charts, heat maps). Support result export (such as PDF, Excel format) to facilitate decision makers to record and share.

[0130] The data storage and management module includes:

[0131] Unified storage of asset usage records, prediction models and optimization solutions to ensure data security and traceability. Adopt relational databases (such as MySQL) and distributed storage technology to support large-scale data management; provide data encryption, permission management and backup mechanisms to ensure data security.

[0132] The system management module includes:

[0133] Provides background management functions for system operation, including user management, permission setting and log recording. Supports hierarchical permission management, with different users (such as administrators and analysts) accessing different functions. Provides log tracking function to record user operations and data change history for easy auditing and accountability.

[0134] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above-described implementation modes. Within the scope of the present invention and the protection scope of the claims in this field, other forms of embodiments all belong to the protection scope of the present invention.

Claims

1. A method and system for analyzing asset usage based on intelligent prediction and optimization algorithm, characterized in that: The method comprises the following steps: S1. Obtain asset usage records and inventory information from the asset management system; S2. Obtain purchase information and depreciation information from the financial system; S3. Based on the asset usage records obtained in step S1 and the procurement information obtained in step S2, use the time series model and regression analysis method to predict the future usage and demand trend of the asset; S4. Based on the prediction result of step S3, the inventory information obtained in step S1 and the depreciation information obtained in step S2, a procurement and configuration plan is generated in combination with an optimization algorithm; S5. Use visualization tools to display the prediction and optimization results of steps S3 and S4 to provide support for asset management and procurement decisions.

2. The asset usage analysis method and system based on intelligent prediction and optimization algorithm according to claim 1, characterized in that: The specific method of obtaining asset usage records and inventory information in step S1 is: S11. Connect to the asset management system and automatically obtain asset usage records (including usage frequency, usage duration, failure rate, etc.) data set S1 and real-time inventory information (current inventory, assets in transit, usage status) data set S2 through API or database query; S12. Based on the asset usage record dataset S1 and the real-time inventory information dataset S2, remove missing values ​​and unreasonable values, unify the format (such as date, amount and other fields), and associate various asset records through primary keys (such as asset numbers) to ensure data consistency, thus forming the asset usage dataset S3.

3. The asset usage analysis method and system based on intelligent prediction and optimization algorithm according to claim 1, characterized in that: The specific method of obtaining the purchase information and depreciation information in step S2 is: S21. Import the historical purchase order data set S4 from the financial system, including the purchase date, purchase quantity, supplier information and purchase amount, and obtain the depreciation policy (such as straight-line method, double declining balance method) and specific depreciation details (such as net asset value, accumulated depreciation) data set S5; S22. Calculate the residual value of the asset according to the depreciation algorithm to form the asset residual value data set S6. Use the asset number or name to associate the historical purchase order data set S4, the asset usage data set S3 and the asset residual value data set S6 to form the asset operation information data set S7.

4. The asset usage analysis method and system based on intelligent prediction and optimization algorithm according to claim 1, characterized in that: The specific method of using the time series model and regression analysis method to predict the future usage and demand trend of assets in step S3 is: S31. Use the historical asset usage dataset S3 to build a time series difference autoregressive moving average ARIMA model for short-term forecasting. The three key parameters of the ARIMA model are: (p, d, q), where p (autoregressive order): depends on the data values ​​of the previous p time points; d (number of differences): the number of difference operations required to stabilize the data; q (moving average order): depends on the error value of the previous q time points. Use ADF (Augmented Dickey-Fuller) or KPSS test to determine whether the time series is stable. If it is not stable, a difference operation is required: AND' t =And t -AND t-1 Then, draw the ACF (autocorrelation function, used to identify the moving average order q) and PACF (partial autocorrelation function, used to identify the autoregressive order p), and determine the p and q values ​​according to the cutoff point. Use actual data to compare the model prediction results and calculate the evaluation indicators: (Mean Square Error (MSE)) (Mean Absolute Error (MAE)) Based on the forecast results, determine the future asset demand and usage frequency. The ARIMA model achieves accurate forecasting of asset usage through the stabilization of historical data, autocorrelation analysis and parameter optimization. Combined with residual analysis and performance evaluation, the reliability of the model and the availability of forecast results can be ensured, providing a scientific basis for asset usage planning and procurement optimization. S32. Use the historical asset usage dataset S3 to build a time series long short-term memory network LSTM model for short-term prediction. First, the data needs to be normalized. In order to accelerate model convergence, the asset usage is normalized: Convert the time series into the input format of the LSTM model (sequence-prediction problem): input-historical data of fixed window length (such as usage in the past 30 days), output-predicted future usage (such as the 31st day). Divide the data into a training set (for model training (about 70% to 80%)), a validation set (to adjust model parameters (about 10% to 15%)), and a test set (to evaluate model performance (about 10%-15%)). Use the training data to fit the model, and use the validation data to monitor overfitting. The monitoring indicator loss function is: By processing historical asset usage records through the LSTM model, we can effectively capture long-term dependencies and complex patterns in time series and generate high-precision future usage forecasts. This provides a scientific basis for asset usage planning and procurement optimization, and improves resource utilization efficiency. S33. Use the historical asset usage dataset S3 to build a forecasting model to quantify the impact of various factors on asset demand and provide support for future asset usage planning. First, the data needs to be preprocessed. Interpolation or mean filling is used for missing data, the Z-score method is used to detect and remove outliers, and numerical data is standardized to avoid the impact of feature magnitude differences on the model: Where μ is the mean and σ is the standard deviation. Determine the predictor variables (X): time factors (such as weeks, months); asset categories (such as equipment, tools); frequency of use, inventory status. Determine the target variable (Y): future usage of the asset (such as demand in the next week or the next quarter). Then, different regression analysis methods are used according to different application scenario requirements. When there is a linear relationship between asset demand and influencing factors, such as demand gradually increasing over time, linear regression is suitable: Y=β0+β1X1+β2X2+…+β n X n +∈ Where, Y: predicted asset usage; X: i : Influencing factors (such as time, inventory status); β i : regression coefficient, quantifying the impact of each feature on asset demand; ∈: error term. When asset demand is affected by multiple factors (such as seasonal fluctuations and inventory levels), it is more effective to use multiple linear regression: When the relationship between asset demand and some factors is nonlinear (such as seasonal fluctuations), polynomial regression is applicable: In high-dimensional data (such as those containing multiple asset types and external factors), select key features, use ridge regression and LASSO regression, and introduce regularization terms to solve multicollinearity or overfitting problems: Ridge Regression:Minimize||Y-Xβ|| 2 +λ||β|| 2 LASSO:Minimize||Y-Xβ|| 2 +λ||β||1 5. The asset usage analysis method and system based on intelligent prediction and optimization algorithm according to claim 1 is characterized in that: The specific method of generating the procurement and configuration plan by the optimization algorithm in step S4 is: Input variables: forecasted asset usage demand, current inventory status, and asset depreciation; Optimization goal: Minimize total cost (purchase cost + maintenance cost + depreciation loss) to meet demand and inventory security constraints; Generate a procurement plan: determine procurement time, procurement quantity and preferred suppliers; Optimize asset allocation: Dynamically adjust the allocation of assets among departments to reduce idleness or shortages. Linear Programming Model: Minimize:C total =C purchase +C maintenance +C depreciation 6. The asset usage analysis method and system based on intelligent prediction and optimization algorithm according to claim 1 is characterized in that: The specific method of displaying the prediction and optimization results by using visualization tools in step S5 to support decision making is as follows: Data visualization: Use charts (such as line charts, bar charts, and heat maps) to display predicted asset usage and demand trends, and display optimized procurement plans and asset allocation plans (such as optimal procurement time and quantity); Dynamic interaction: Provides a user interaction interface, supports conditional filtering (such as time period, asset type), and supports data export (such as generating procurement reports).

Images