An engineering material procurement optimization method and system considering weather and price fluctuations

Abstract

The application provides an engineering material procurement optimization method and system considering weather and price fluctuation, belongs to the field of engineering material management and supply chain optimization, and comprises the following steps: obtaining engineering material historical procurement data, supplier quotation data, weather data, engineering progress data and inventory data to obtain a standard data set; constructing a nonlinear price delivery date function according to the standard data set to obtain the price delivery date trade-off relationship of different materials at different suppliers, and obtaining construction period fluctuation data and material demand fluctuation data by analyzing the construction period fluctuation characteristics and evaluating the material demand fluctuation; finally, a multi-objective optimization function is constructed according to the obtained data, and the multi-objective optimization function is solved by using an NSGA-II algorithm with a rolling window mechanism to generate an engineering material procurement optimization decision. The application solves the problems of the existing engineering material management technology in the aspects of price delivery date trade-off modeling, dynamic progress adaptability, inventory and material shortage risk collaborative control and the like.

Description

Technical Field

[0001] This invention belongs to the field of engineering material management and supply chain optimization, and in particular relates to an optimization method and system for engineering material procurement that takes into account weather and price fluctuations. Background Technology

[0002] In engineering construction, especially in large-scale infrastructure projects such as dams, canals, highways, ports, and railways, the procurement and inventory management of engineering materials is a crucial link in ensuring the smooth progress of the project. Currently, the supply model for engineering materials mainly relies on fixed-price contracts, periodic order placement and delivery, and designated on-site storage, combined with manual judgment and experience-based replenishment decisions based on the project construction plan. This traditional model faces multiple challenges in practice:

[0003] On the one hand, the prices of construction materials are highly volatile. Taking bulk building materials such as manufactured sand, cement, and steel as examples, their prices are greatly affected by factors such as market supply and demand, transportation costs, and regional differences. Especially during the rainy season or periods of policy adjustment, price fluctuations are significant, easily leading to concentrated procurement during periods of high prices and increasing overall costs. On the other hand, the progress of large-scale infrastructure projects is highly uncertain. Influencing factors include weather changes (such as heavy rain and typhoons), delays in on-site construction progress, and adjustments to construction organization, resulting in frequent adjustments to actual material usage plans and alternating periods of "material shortages causing work stoppages" and "stockpiling and aging." Existing methods struggle to balance ensuring project completion and controlling inventory costs.

[0004] Furthermore, existing technologies generally overlook the multi-dimensional trade-offs arising from the supply characteristics of low-price, long-delivery-time versus high-price, fast-delivery-time. Suppliers often offer two drastically different supply strategies: one is low-price but long-delivery-time, suitable for planned replenishment; the other is high-price but timely delivery, suitable for emergency procurement. Current systems lack modeling and quantification mechanisms for these price-price trade-offs, leading to one-sided procurement decisions and increasing the risk of ad-hoc procurement and cost overruns.

[0005] In terms of technical implementation, some studies have attempted to introduce basic inventory control models (such as the EOQ model and rolling replenishment model) to cope with volatile demand, but have failed to effectively integrate key variables such as dynamic adjustment of construction plans, price forecasting, and supply responsiveness. At the project site management level, there are also problems such as lack of information transparency, untimely scheduling, and lagging system response, making it difficult to dynamically coordinate procurement and inventory strategies. Summary of the Invention

[0006] To address the aforementioned shortcomings in existing technologies, this invention provides an optimization method and system for engineering material procurement that takes into account weather and price fluctuations. This method solves the significant deficiencies in existing engineering material management technologies regarding price trade-off modeling, dynamic schedule adaptability, and coordinated control of inventory and material shortage risks.

[0007] To achieve the above objectives, the technical solution adopted by this invention is: an optimization method for engineering material procurement that takes into account weather and price fluctuations, comprising the following steps:

[0008] S1. Obtain historical procurement data, supplier quotation data, weather data, project progress data, and inventory data for engineering materials, and perform data cleaning and standardization to obtain a standard dataset;

[0009] S2. Based on the historical procurement data of engineering materials and supplier quotation data in the standard dataset, construct a nonlinear price-delivery function to obtain the price-delivery balance relationship of different materials from different suppliers.

[0010] S3. Based on the weather data, project progress data and inventory data in the standard dataset, analyze the characteristics of project duration fluctuations and assess the fluctuations in material demand to obtain project duration fluctuation data and material demand fluctuation data.

[0011] S4. Based on the price trade-off relationship of different materials with different suppliers, the project duration fluctuation data, and the material demand fluctuation data, construct a multi-objective optimization function, and use the NSGA-II algorithm with rolling window mechanism to solve the multi-objective optimization function and output the Pareto optimal solution set.

[0012] S5. Based on the Pareto optimal solution set, use visualization technology to generate optimization decisions for engineering material procurement.

[0013] Furthermore, S2 specifically includes:

[0014] S201. Use STL decomposition to perform time series decomposition on the material price series in the historical procurement data of engineering materials to obtain the decomposed time series.

[0015] S202. Analyze the decomposed time series using the autocorrelation function, partial autocorrelation function, and unit root test to obtain the stationarity test results of the series.

[0016] S203. Based on the stationarity test results of the sequence, establish a future price prediction model for commodities based on the ARIMA model to obtain the predicted commodity prices.

[0017] S204. Based on the commodity price series and predicted commodity prices, construct a nonlinear price delivery function to obtain the price delivery balance relationship of different commodities from different suppliers.

[0018] The further beneficial effects mentioned above are: by accurately predicting price trends through STL decomposition and ARIMA model, and by quantifying the intrinsic relationship between supplier delivery time and price using nonlinear functions, dynamic and reliable procurement decision-making basis is provided for multi-objective optimization, thereby improving cost control and supply stability.

[0019] Furthermore, the expression for the nonlinear price delivery function is as follows:

[0020]

[0021]

[0022]

[0023]

[0024]

[0025] in, For the first The first predicted path The price of goods at any given time For the predicted path number, In terms of time dimension, For the first Monthly forecast of commodity prices, For the first The price fluctuation disturbance term at time t under the path, Let the disturbance term follow a normal distribution. This is the fluctuation adjustment factor. The standard deviation of the fitted residuals of the nonlinear price delivery function. This represents the upper limit of the price prediction at time t with a 95% confidence level. This represents the lower bound of the price prediction at time t with a 95% confidence level. , and All are fusion coefficients. For the first Monthly forecast of commodity prices, This refers to the latest actual transaction price in the historical price series. This is the arithmetic average of all historical purchase prices.

[0026] The further beneficial effects mentioned above are as follows: the nonlinear price-delivery function accurately portrays the inherent trade-off between price and delivery through exponential relationships, combines multi-path simulation and confidence intervals to quantify market uncertainty, and integrates historical information through fusion strategies, which significantly improves the accuracy of procurement decisions, risk predictability, and strategy robustness.

[0027] Furthermore, S3 specifically includes:

[0028] S301. Calculate the weather impact coefficient based on weather data;

[0029] S302. Based on the project progress data, analyze the project duration fluctuation characteristics by considering the difference between the actual project duration and the planned project duration and the project duration compensation amount, and calculate the project duration fluctuation data.

[0030] S303. Based on the weather impact coefficient, assess the fluctuation of material demand on the inventory data to obtain material demand fluctuation data.

[0031] The further beneficial effects mentioned above are: by quantifying the impact of weather factors on construction efficiency, accurately capturing schedule fluctuations by combining a dynamic progress compensation mechanism, and realizing real-time assessment of material demand based on multi-source data fusion, the prediction accuracy and response capability for demand fluctuations in complex engineering environments are significantly improved.

[0032] Furthermore, the expression for the project duration fluctuation data is as follows:

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040] in, This represents the probability density function value for project duration fluctuations during the initial startup phase. For the initial standardized construction period, This is the initial resource delivery dispersion parameter. Parameters for the initial integration and coordination of personnel. For beta functions, This represents the probability density function value of the construction period fluctuation in the mid-term intensive construction phase. For the actual construction period in the middle stage, The location parameters of the mid-term log-normal distribution. The scaling parameter of the mid-log-normal distribution. The cumulative distribution function of the standard normal distribution. This is the upper limit of the mid-term construction period cutoff interval. This is the lower limit of the mid-term construction period cutoff interval. The planned construction period for the mid-term construction phase, This represents the probability density function value for schedule fluctuations during the later, final stages of the project. For the actual construction period in the later stage, For the number of rework sessions, For the intensity parameters of the later Poisson distribution, For the first Average construction period after each rework For the first Standard deviation of the project timeline after rework This is an approximation of the probability density function for the project duration fluctuation in the system's phase transition zone. For the derivation For the characteristic function, For the actual construction period of the transition zone, For the stability parameters of the Lévy distribution, The skewness parameter of the Lévy distribution. Let Lévy's distribution scale parameter be denoted as . The independent variable of the characteristic function is... For symbolic functions, For gamma function, For the first Daily project schedule delay For the first Daily schedule progress For the first Daily actual progress For the first Daily cumulative uncompensated progress deviation For the summation index, For the first Daily adjusted resource input As the baseline resource input, For resource adjustment coefficients, To compensate for the apportionment of subsequent days, This represents the number of days between the current date and the date of the delayed occurrence.

[0041] The further beneficial effects mentioned above are as follows: by establishing beta distribution, truncated log-normal distribution, Poisson-Gaussian mixture distribution and Lévy distribution in stages, the characteristics of schedule fluctuation at different construction stages are characterized. Combined with the exponential decay compensation mechanism, dynamic resource adjustment is realized, which improves the accuracy of schedule prediction and the adaptability of resource scheduling.

[0042] Furthermore, the expression for the fluctuation data of material demand is as follows:

[0043]

[0044]

[0045]

[0046]

[0047]

[0048] in, For the first Daily final demand for supplies, i.e., data on fluctuations in supply demand. For the first Daily demand for basic supplies For the first Daily material demand compensation amount Let be the random fluctuation factor of demand on day t. For the first Daily demand for raw materials For the first Daily weather impact coefficient, For the first Daily rain resistance coefficient, For the first Daily rainfall, For the first Daily drag coefficient, For the first Daily average wind speed For the first Maximum daily gust wind speed.

[0049] The further beneficial effects mentioned above are: by quantifying the impact of rainfall and wind speed on construction efficiency to establish a weather impact coefficient, and combining it with the exponential decay compensation mechanism for schedule lag and random fluctuation factors, dynamic and accurate prediction of material demand can be achieved, significantly improving the adaptability and accuracy of demand planning in complex environments.

[0050] Furthermore, the expression for the multi-objective optimization function in S4 is as follows:

[0051]

[0052]

[0053]

[0054]

[0055] in, The total cost objective is optimized for multiple objectives. For material category indexing, This represents the total number of categories of materials required for the project. In terms of time dimension, To optimize the total number of days in the cycle, For the first Category of materials Daily purchase volume For the first Category of materials Daily purchase unit price, For the first The unit daily storage cost of this type of material, For the first Category of materials Daily inventory For the first Fixed cost per order for this type of material. For the first Category of materials Daily purchasing decision variables The material shortage risk objective is optimized for multiple objectives. For the first Category of materials Actual daily demand Safety stock coverage target optimized for multiple objectives. For the first Category of materials Daily safety stock level Heuristic inventory risk objectives for multi-objective optimization For the first The aging risk coefficient of high-risk materials This is the inventory aging and decay coefficient. For the first Category of materials Daily inventory backlog days.

[0056] The further beneficial effects mentioned above are as follows: By simultaneously balancing procurement costs, material shortage risks, safety stock coverage, and material aging risks, the multi-objective optimization model breaks through the limitations of traditional single-objective optimization, achieves synergistic optimization of cost control and supply chain resilience, and improves the economy and reliability of engineering material management.

[0057] Furthermore, the multi-objective optimization function is configured with constraints on storage capacity, procurement lead time, minimum order quantity, and safety stock minimum.

[0058] The expression for the storage capacity constraint is as follows:

[0059]

[0060] in, For the first Total daily storage occupancy of all materials For material category indexing, In terms of time dimension, For the first The percentage of each type of material by unit volume or unit weight. Total storage capacity To define the scope of application;

[0061] The expression for the procurement lead time constraint is as follows:

[0062]

[0063] in, For the first Category of materials Daily inventory For the first Category of materials Daily inventory For the first Category of materials Daily purchase volume For the time dimension of purchasing orders, it means The supplies ordered today are Same-day delivery For the first Lead time for procurement of such materials For the first Category of materials Actual daily demand To define the scope of application;

[0064] The expression for the minimum order quantity constraint is as follows:

[0065]

[0066] in, For the first Category of materials Daily purchase volume This is the upper limit of the maximum order quantity for supplies. For the first Minimum order quantity for this type of goods. For the first Category of materials Daily purchasing decision variables For the first The upper limit of the maximum order quantity for this type of material;

[0067] The expression for the safety stock lower limit constraint is as follows:

[0068]

[0069] in, For the first Category of materials Daily safety stock level Safety stock factor For the first Category of materials Actual daily demand To constrain the scope of application.

[0070] The further beneficial effects mentioned above are as follows: the multi-objective optimization function is set with constraints on storage capacity, procurement lead time, minimum order quantity, and safety stock lower limit. Storage capacity constraint ensures storage feasibility, procurement lead time constraint ensures timely supply, minimum order quantity constraint optimizes procurement economy, and safety stock lower limit constraint prevents material shortage risk. Together, they construct an optimization decision space that conforms to actual operating conditions.

[0071] This invention also provides an engineering material procurement optimization system that takes into account weather and price fluctuations, comprising:

[0072] The data acquisition and preprocessing module is used to acquire historical procurement data of engineering materials, supplier quotation data, weather data, project progress data and inventory data, and to perform data cleaning and standardization to obtain a standard dataset.

[0073] The supply characteristic analysis module is used to analyze the characteristics of project schedule fluctuations and assess the fluctuations in material demand based on weather data, project progress data and inventory data in the standard dataset, and obtain the price trade-off relationship of different materials from different suppliers.

[0074] The demand fluctuation assessment module is used to obtain project schedule fluctuation data and material demand fluctuation data by analyzing the characteristics of project schedule fluctuation and assessing the fluctuation of material demand based on weather data, project progress data and inventory data in the standard dataset.

[0075] The multi-objective optimization solution module is used to construct a multi-objective optimization function based on the trade-off relationship between different materials and different suppliers, the fluctuation data of construction period and the fluctuation data of material demand. The module then uses the NSGA-II algorithm with a rolling window mechanism to solve the multi-objective optimization function and output the Pareto optimal solution set.

[0076] The decision generation and visualization module is used to generate optimal decisions for engineering material procurement based on the Pareto optimal solution set using visualization technology.

[0077] The beneficial effects of this invention are:

[0078] (1) Under the premise of meeting the requirements of engineering construction schedule, this invention effectively reduces the total procurement cost by constructing a multi-dimensional optimization model of cost-delivery-risk. Compared with the traditional single cost minimization strategy, this invention can reduce the average procurement cost in a large-scale project simulation, while reducing sudden high-price procurement behavior caused by delivery uncertainty. This invention uses the NSGA-II algorithm for automatic optimization to achieve dynamic matching of procurement batch and delivery strategy, reducing waste caused by procurement misjudgment dominated by human experience.

[0079] (2) This invention constructs a mathematical model covering five dimensions of objectives, and introduces a composite objective of maximizing safety stock coverage and minimizing heuristic inventory risk. It can take into account both short-term inventory security and long-term material aging control, and optimize turnover rate while ensuring safety. The storage capacity constraint of this invention can dynamically identify materials that are prone to aging and accumulation, and strategically avoid them by selecting strategies based on actual delivery dates, thereby ensuring the health and sustainability of material flow.

[0080] (3) This invention can control the risk of material shortage through a multi-objective optimization function, reduce the risk of project delays caused by material supply disruptions, improve the stability of project performance, and enhance the social reputation of the project. It promotes green, low-carbon and resource-saving practices in the construction process and reduces unnecessary material stockpiling, over-purchasing and duplicate transportation through intelligent inventory strategies.

[0081] (4) The engineering material procurement and inventory optimization method proposed in this invention has significant technical advantages in reducing overall costs, enhancing supply guarantee capabilities, and controlling inventory risks, and has good practical application prospects and promotion value. Attached Figure Description

[0082] Figure 1 This is an optimization method for engineering material procurement that takes into account weather and price fluctuations. Detailed Implementation

[0083] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0084] Example 1

[0085] like Figure 1 The diagram illustrates an optimization method for engineering material procurement that takes into account weather and price fluctuations, comprising the following steps:

[0086] S1. Obtain historical procurement data, supplier quotation data, weather data, project progress data, and inventory data for engineering materials, and perform data cleaning and standardization to obtain a standard dataset;

[0087] S2. Based on the historical procurement data of engineering materials and supplier quotation data in the standard dataset, construct a nonlinear price-delivery function to obtain the price-delivery balance relationship of different materials from different suppliers.

[0088] S3. Based on the weather data, project progress data and inventory data in the standard dataset, analyze the characteristics of project duration fluctuations and assess the fluctuations in material demand to obtain project duration fluctuation data and material demand fluctuation data.

[0089] S4. Based on the price trade-off relationship of different materials with different suppliers, the project duration fluctuation data, and the material demand fluctuation data, construct a multi-objective optimization function, and use the NSGA-II algorithm with rolling window mechanism to solve the multi-objective optimization function and output the Pareto optimal solution set.

[0090] S5. Based on the Pareto optimal solution set, use visualization technology to generate optimization decisions for engineering material procurement.

[0091] This invention provides an optimized procurement method for engineering materials that considers weather and price fluctuations. It addresses the shortcomings of traditional methods, which suffer from insufficient adaptability and decision-making lag when facing the dual uncertainties of changing weather conditions and fluctuating market prices. These problems lead to high procurement costs, uncontrollable material shortage risks, and low inventory turnover efficiency. This invention utilizes multi-source data fusion technology to integrate historical procurement data, supplier quotation data, real-time weather data, project progress data, and inventory data to establish a complete standardized dataset, providing a reliable data foundation for subsequent analysis. Based on this, a nonlinear price-delivery function is constructed to quantify the inherent trade-offs between price and delivery cycle for different suppliers, providing a scientific basis for procurement strategy formulation. Simultaneously, this invention analyzes the impact mechanism of weather factors on construction efficiency, and combines the deviation characteristics between actual and planned project progress to output data on schedule fluctuations and material demand fluctuations. This enables the material procurement plan to respond promptly to changes in construction conditions, improving the accuracy of material demand forecasting. This invention designs a multi-objective optimization model that comprehensively considers procurement costs, material shortage risks, safety stock coverage, and material aging risks. It uses the NSGA-II intelligent algorithm with a rolling window mechanism for dynamic solution, which can balance various engineering requirements, achieve optimal material procurement decisions, and transform them into intuitive and easy-to-understand procurement decision schemes through visualization technology, thereby improving decision-making efficiency and operability.

[0092] In one embodiment of the present invention, S1, historical procurement data, supplier quotation data, weather data, project progress data, and inventory data of engineering materials are acquired, and the data is cleaned and standardized to obtain a standard dataset. Barcode identification devices can be used to identify the materials, and barcode scanning devices can be used to recognize the materials, acquiring historical procurement data, supplier quotation data, and inventory data of engineering materials. When materials are put into storage, the barcode scanning device reads the barcode information and automatically associates it with data such as the purchase order number and arrival time, uploading it to the central server in real time. When materials are put out of storage, the barcode is scanned to verify information such as the receiving department, construction location, and quantity received, and the inventory data is updated synchronously to ensure that the inventory dynamics are consistent with actual consumption, avoiding the risk of discrepancies between accounts and actual stock. Simultaneously, an industrial touch screen all-in-one machine can be configured and deployed in the project management center, connected to a laptop computer for use by on-site management personnel. This allows for real-time access to weather data and project progress data, as well as real-time retrieval of material data and viewing of procurement optimization plans, adapting to the needs of rapid querying and immediate decision-making at the project site.

[0093] In one embodiment of the present invention, addressing the problems of unpredictable price fluctuations and difficulty in quantifying the supplier price-price trade-off in traditional procurement, S2 in the present invention achieves accurate prediction of material prices and quantitative characterization of supplier delivery time characteristics through time series analysis and nonlinear function construction, providing key data support for optimizing procurement decisions. S2 specifically includes:

[0094] S201. Use STL decomposition to perform time series decomposition on the material price series in the historical procurement data of engineering materials to obtain the decomposed time series.

[0095] S202. Analyze the decomposed time series using the autocorrelation function, partial autocorrelation function, and unit root test to obtain the stationarity test results of the series.

[0096] S203. Based on the stationarity test results of the sequence, establish a future price prediction model for commodities based on the ARIMA model to obtain the predicted commodity prices.

[0097] S204. Based on the commodity price series and predicted commodity prices, construct a nonlinear price delivery function to obtain the price delivery balance relationship of different commodities from different suppliers; the expression of the nonlinear price delivery function is as follows:

[0098]

[0099]

[0100]

[0101]

[0102] in, For the first The first predicted path The price of goods at any given time For the predicted path number, The time dimension is used, with the unit being months. For the first Monthly forecast of commodity prices, For the first The price fluctuation disturbance term at time t under the given path. Let the disturbance term follow a normal distribution. This is the fluctuation adjustment factor. The standard deviation of the fitted residuals of the price-delivery function reflects the deviation between the supplier's quotation and the fitted function. This represents the upper limit of the price prediction at time t with a 95% confidence level. This represents the lower bound of the price prediction at time t with a 95% confidence level. , and All are fusion coefficients. For the first Monthly forecast of commodity prices, This refers to the latest actual transaction price in the historical price series. The arithmetic mean of all historical procurement prices is used. By establishing a nonlinear price-delivery function, the price-delivery trade-off relationship of different materials from different suppliers can be obtained, providing input parameters for subsequent multi-objective optimization functions. This facilitates meeting engineering requirements while satisfying cost control, avoiding material supply delays caused by pursuing low prices, or cost increases caused by pursuing rapid material supply.

[0103] In one embodiment of the present invention, addressing the problem that traditional material demand forecasting neglects the impact of weather on project progress, S3 of the present invention establishes an evaluation mechanism for weather impact coefficients and project duration fluctuation data to quantify the correlation between project construction efficiency and material consumption, obtaining material fluctuation data and project duration fluctuation data. This enables material demand forecasting to dynamically adapt to weather changes and project progress, improving the matching accuracy between material procurement plans and actual project needs. S3 specifically includes:

[0104] S301. Calculate the weather impact coefficient based on weather data;

[0105] S302. Based on the project progress data, and considering the difference between the actual and planned project durations, as well as the project duration compensation amount, the project duration fluctuation characteristics are analyzed, and the project duration fluctuation data is calculated. The expression for the project duration fluctuation data is as follows:

[0106]

[0107]

[0108]

[0109]

[0110]

[0111]

[0112]

[0113] in, This represents the probability density function value for project duration fluctuations during the initial startup phase. For the initial standardized construction period, This is the initial resource delivery dispersion parameter. Parameters for the initial integration and coordination of personnel. For beta functions, This represents the probability density function value of the construction period fluctuation in the mid-term intensive construction phase. For the actual construction period in the middle stage, The location parameters of the mid-term log-normal distribution. The scaling parameter of the mid-log-normal distribution. The cumulative distribution function of the standard normal distribution. This is the upper limit of the mid-term construction period cutoff interval. This is the lower limit of the mid-term construction period cutoff interval. The planned construction period for the mid-term construction phase; This represents the probability density function value for schedule fluctuations during the later, final stages of the project. The actual construction period is in days. For the number of rework sessions, For the intensity parameters of the later Poisson distribution, For the first Average construction period after each rework This represents the average time spent on a single rework. For the first Standard deviation of the project timeline after rework This is an approximation of the probability density function for the project duration fluctuation in the system's phase transition zone. For the derivation For the characteristic function, For the actual construction period of the transition zone, For the stability parameters of the Lévy distribution, The skewness parameter of the Lévy distribution. Let Lévy's distribution scale parameter be denoted as . The independent variable of the characteristic function is... For a sign function, when When it is 1, When it is -1, For gamma function, For the first Daily project schedule delay For the first Daily schedule progress For the first Daily actual progress For the first Daily cumulative uncompensated progress deviation To find the sum index, iterate from day 1 to day t. For the first Daily adjusted resource input As the baseline resource input, This is a resource adjustment coefficient; To compensate for the apportionment of subsequent days, This represents the number of days between the current date and the date of the delayed occurrence.

[0114] S303. Based on the weather impact coefficient, assess the fluctuation of material demand on the inventory data to obtain material demand fluctuation data; the expression for the material demand fluctuation data is as follows:

[0115]

[0116]

[0117]

[0118]

[0119]

[0120] in, For the first Daily final demand for supplies, i.e., data on fluctuations in supply demand. For the first Daily demand for basic supplies For the first Daily material demand compensation amount Let be the random fluctuation factor of demand on day t. For the first Daily demand for raw materials For the first Daily weather impact coefficient, For the first Daily rain resistance coefficient, For the first Daily rainfall, For the first Daily drag coefficient, For the first Daily average wind speed For the first Maximum daily gust wind speed.

[0121] In one embodiment of the present invention, S4, based on the price trade-off relationship of different materials from different suppliers, construction period fluctuation data, and material demand fluctuation data, a multi-objective optimization function is constructed, and the NSGA-II algorithm with a rolling window mechanism is used to solve the multi-objective optimization function to output the Pareto optimal solution set; wherein, the expression of the multi-objective optimization function in S4 is as follows:

[0122]

[0123]

[0124]

[0125]

[0126] in, The total cost objective is optimized for multiple objectives. Index for material categories , This represents the total number of categories of materials required for the project. For the time dimension , To optimize the total number of days in the cycle, For the first Category of materials Daily purchase volume For the first Category of materials Daily purchase unit price, For the first The unit daily storage cost of this type of material, For the first Category of materials Daily inventory For the first Fixed cost per order for this type of material. For the first Category of materials Daily purchasing decision variables Indicates the first Japan to the first Procurement of such materials Indicates no purchase. The material shortage risk objective is optimized for multiple objectives. For the first Category of materials Actual daily demand Safety stock coverage target optimized for multiple objectives. For the first Category of materials Daily safety stock level Heuristic inventory risk objectives for multi-objective optimization For the first The aging risk coefficient of high-risk materials This is the inventory aging and decay coefficient. For the first Category of materials Daily inventory backlog days.

[0127] The multi-objective optimization function of this invention not only considers minimizing total cost J1, but also simultaneously introduces minimizing material shortage risk J2, maximizing safety stock coverage J3, and minimizing heuristic inventory risk J4, forming a multi-objective collaborative optimization mechanism. This mechanism can more realistically reflect the complexity of multi-objective trade-offs in actual engineering operations and is significantly superior to traditional single-objective methods that only pursue the lowest cost or lowest inventory. Furthermore, addressing the issues of environmental uncertainty and dynamic changes in engineering material demand in engineering projects, this invention introduces a rolling optimization mechanism, employing a strategy of forward decision-making and rolling updates. Only some recent decisions are executed in each optimization cycle, retaining the flexibility to adjust for the future. This significantly enhances the robustness and implementability of the model, outperforming existing static optimization models and making it suitable for real-time scheduling and decision reconfiguration needs in actual engineering projects. This invention utilizes the NSGA-II algorithm with a rolling window mechanism to solve problems, significantly improving computational efficiency and solution set diversity. It addresses the shortcomings of traditional multi-objective optimization algorithms in terms of solution distribution and computation time, and makes adaptive adjustments based on the problem structure. Compared to traditional algorithms, the NSGA-II algorithm with a rolling window mechanism in this invention is superior in terms of convergence speed, uniformity of Pareto front distribution, and coverage of non-dominated solutions. It also has better scalability and solution capabilities, especially in large and medium-sized engineering material scheduling scenarios.

[0128] In a specific embodiment of the present invention, the multi-objective optimization function is configured with constraints such as storage capacity constraints, procurement lead time constraints, minimum order quantity constraints, and safety stock lower limit constraints.

[0129] The expression for the storage capacity constraint is as follows:

[0130]

[0131] in, For the first Total daily storage occupancy of all materials For material category indexing, The time dimension is used, with the unit being days. For the first The percentage of each type of material by unit volume or unit weight. Total storage capacity To define the scope of application;

[0132] The expression for the procurement lead time constraint is as follows:

[0133]

[0134] in, For the first Category of materials Daily inventory For the first Category of materials Daily inventory For the first Category of materials Daily purchase volume For the time dimension of purchasing orders, it means The supplies ordered today are Same-day delivery For the first Lead time for procurement of such materials For the first Category of materials Actual daily demand The scope of application is indicated by the following symbols: procurement lead time constraint, minimum order quantity constraint, and safety stock minimum constraint apply to all material categories and all times.

[0135] The expression for the minimum order quantity constraint is as follows:

[0136]

[0137] in, For the first Category of materials Daily purchase volume This is the upper limit of the maximum order quantity for supplies. For the first Minimum order quantity for this type of goods. For the first Category of materials Daily purchasing decision variables Indicates the first Daily procurement Type of supplies, Indicates no purchase. For the first The upper limit of the maximum order quantity for this type of material;

[0138] The expression for the safety stock lower limit constraint is as follows:

[0139]

[0140] in, For the first Category of materials Daily safety stock level Safety stock factor For the first Category of materials Actual daily demand The scope of application is indicated by the constraint that the procurement lead time constraint, minimum order quantity constraint, and safety stock lower limit constraint apply to all material categories and all times.

[0141] In one embodiment of the present invention, S5, based on the Pareto optimal solution set, an optimization decision for the procurement of engineering materials is generated using visualization technology;

[0142] The specific process for calculating the Pareto optimal solution set is as follows:

[0143] A1. Custom encoding strategy:

[0144] The procurement-inventory decision variables obtained by solving the multi-objective optimization function are encoded using a 20-bit binary chromosome. The chromosome structure is as follows:

[0145]

[0146] The encoding rules are defined as follows:

[0147] 1st-5th place ( ): Whether to make a procurement decision in advance (0-1 variable), corresponding to 5 types of core materials, such as cement, steel bars, sand and gravel, etc. Indicates the first Choose to purchase such supplies in advance. This indicates that no advance purchases will be made. );

[0148] 6th-15th positions ( ): Procurement method decision (0-1 variable), corresponding to 10 major categories of materials. Indicates the first For this type of material, choose the fast delivery method. This indicates that the normal delivery method has been selected. );

[0149] 16th-20th place ( ): Storage tank capacity selection (discrete variable), corresponding to 5 storage tanks. The binary value of is represented by the decimal value after conversion. The capacity levels of each tank, such as 000→100m³, 001→200m³, ..., 111→700m³, are actually determined by calibration based on engineering inventory requirements. ).

[0150] A2. Population Evolution and Non-Dominant Sequencing:

[0151] Initial population generation: Randomly generated One chromosome, We can choose 100, which represents the population size, to form the initial solution set. ;

[0152] Iterative Evolution: For the first Generation population Through crossover, crossover probability Mutation, mutation probability Operation to generate offspring population , merged into ;

[0153] Non-dominated sorting: Each chromosome Calculate its corresponding multi-objective function value. According to the non-dominant relationship Divided into multiple non-dominated layers ,in This is the first level of non-dominated solution, meaning that no other solution is superior to this level solution on all objectives;

[0154] Crowding calculation: for each non-dominated layer Calculate each solution Crowding The formula is:

[0155]

[0156] in, For the first The objective function value, correspond arrive , For targets in the same layer After sorting Adjacent solutions, For the goal exist Maximum / minimum values ​​in;

[0157] Population selection: Solutions are selected from high to low priority based on the non-dominated layer, and within the same layer, they are selected from high to low crowding, until the population size is restored to normal. , obtained the Generation population ;

[0158] Termination condition: When the number of iterations reaches... Reaching the maximum number of iterations At that time, evolution ceases, and the final population is selected. The first-level non-dominated solution is used as a candidate Pareto solution.

[0159] A3. Expression for the Pareto optimal solution set:

[0160] The final output Pareto optimal solution set Defined as: And it satisfies the solution set size constraint: , Representing the solution set The number of optimized solutions;

[0161] The dominance relationship is defined as follows:

[0162] If for all targets , ,Right now To minimize the target or ,Right now To maximize the objective, and at least one objective exists. Make or Then it is called Dominate , recorded as ;

[0163] like or For all If established, it is called weak dominance , recorded as .

[0164] A4. Output of the optimized solution corresponding to the solution set:

[0165] Pareto optimal solution set Each chromosome Decoded into specific procurement-inventory optimization solutions :

[0166] in, For the first Category of materials The optimal daily purchase quantity is calculated by combining the "whether to purchase in advance" code with the price-delivery function. For the first Category of materials The optimal procurement method for each day is fast / standard delivery, which is directly decoded from the procurement method code. For the first The first storage tank The optimal daily capacity configuration is determined by encoding and decoding the inventory tank capacity; finally, visualization technology is used to transform it into intuitive charts, showing the performance of the procurement-inventory optimization solution in terms of cost, inventory risk, and material shortage risk.

[0167] Example 2

[0168] This invention provides an optimization system for engineering material procurement that takes into account weather and price fluctuations, comprising:

[0169] The data acquisition and preprocessing module is used to acquire historical procurement data of engineering materials, supplier quotation data, weather data, project progress data, and inventory data, and to perform data cleaning and standardization to obtain a standard dataset. In a specific embodiment of the invention, the historical procurement data of engineering materials includes supplier name, material code, quotation amount, actual delivery days, purchase quantity, current supplier quotation, and quotation schemes corresponding to different delivery days. Simultaneously, for inventory data, it can also perform inventory aging risk assessment: a quantitative model of inventory aging loss cost is constructed, calculating aging loss cost = (inventory turnover days × quality degradation coefficient × material unit price) + (inventory turnover days × unit warehousing cost) + (inventory batch quantity × management labor cost), quantitatively assessing the inventory aging risk of various materials. For example, the aging loss amount is automatically calculated for cement stored for more than 45 days, and the optimization model penalizes the scheme of "purchasing large quantities in advance leading to inventory backlog" to suppress high inventory risk.

[0170] The supply characteristic analysis module is used to analyze the characteristics of project schedule fluctuations and assess the fluctuations in material demand based on weather data, project progress data and inventory data in the standard dataset, and obtain the price trade-off relationship of different materials from different suppliers.

[0171] The demand fluctuation assessment module is used to analyze the characteristics of project schedule fluctuations and assess material demand fluctuations based on weather data, project progress data, and inventory data in a standard dataset, thereby obtaining project schedule fluctuation data and material demand fluctuation data. In a specific embodiment of the present invention...

[0172] The multi-objective optimization solution module is used to construct a multi-objective optimization function based on the price trade-off relationship of different materials from different suppliers, project duration fluctuation data, and material demand fluctuation data. It then uses the NSGA-II algorithm with a rolling window mechanism to solve the multi-objective optimization function and outputs a Pareto optimal solution set. In a specific embodiment of this invention, the input parameters of the multi-objective optimization function include: price trade-off relationship and data from project duration fluctuation data and material demand fluctuation data, such as the engineering construction plan (including the start / completion time of each process and the material demand list), storage capacity, preset material shortage days threshold, and safety stock coverage target value.

[0173] A four-objective optimization model is constructed, with objective functions including: minimizing total cost = procurement cost + warehousing cost + ordering cost; minimizing material shortage risk = number of days of material shortage / total construction days × 100%; maximizing safety stock coverage = safety stock days / demand cycle days × 100%; and minimizing inventory aging risk = aging loss cost / total procurement cost × 100%. Three constraints are also set: materials must arrive before construction, average daily inventory ≤ warehousing capacity, and total number of days of material shortage ≤ threshold. This transforms the core requirements of engineering material procurement and inventory management into a mathematical model, providing a standardized "objective-constraint" framework for subsequent algorithm solving, ensuring that the optimization direction aligns with the actual needs of the project.

[0174] The decision generation and visualization module is used to generate optimal decisions for engineering material procurement based on the Pareto optimal solution set using visualization technology.

[0175] The beneficial effects of this invention are as follows: This invention establishes a standardized dataset through multi-source data fusion and constructs a nonlinear price-delivery function to quantify the trade-off between supplier material prices and delivery times. Simultaneously, it establishes a dynamic material demand forecast that considers weather factors and construction period fluctuations, enabling material procurement plans to adapt to changes in construction weather conditions and project progress in real time. By establishing a multi-objective optimization model and employing an intelligent algorithm with a rolling window, it simultaneously optimizes objectives such as procurement costs, material shortage risk, inventory coverage, and material aging, significantly reducing procurement costs, effectively controlling stockout risks, improving inventory turnover efficiency, and enhancing supply chain resilience. Finally, it outputs the optimal procurement plan through visualization technology. This invention improves the scientific nature and environmental engineering adaptability of material procurement decisions, providing a reliable material procurement plan for complex engineering environments and project schedules.

Claims

1. A method for optimizing the procurement of engineering materials considering weather and price fluctuations, characterized in that, Includes the following steps: S1. Obtain historical procurement data, supplier quotation data, weather data, project progress data, and inventory data for engineering materials, and perform data cleaning and standardization to obtain a standard dataset; S2. Based on the historical procurement data of engineering materials and supplier quotation data in the standard dataset, construct a nonlinear price-delivery function to obtain the price-delivery balance relationship of different materials from different suppliers. S3. Based on the weather data, project progress data and inventory data in the standard dataset, analyze the characteristics of project duration fluctuations and assess the fluctuations in material demand to obtain project duration fluctuation data and material demand fluctuation data. S4. Based on the price trade-off relationship of different materials with different suppliers, the project duration fluctuation data, and the material demand fluctuation data, construct a multi-objective optimization function, and use the NSGA-II algorithm with rolling window mechanism to solve the multi-objective optimization function and output the Pareto optimal solution set. S5. Based on the Pareto optimal solution set, use visualization technology to generate optimization decisions for engineering material procurement; S3 specifically includes: S301. Calculate the weather impact coefficient based on weather data; S302. Based on the project progress data, analyze the project duration fluctuation characteristics by considering the difference between the actual project duration and the planned project duration and the project duration compensation amount, and calculate the project duration fluctuation data. S303. Based on the weather impact coefficient, assess the fluctuation of material demand on the inventory data to obtain material demand fluctuation data. The expression for the project duration fluctuation data is as follows: in, This represents the probability density function value for project duration fluctuations during the initial startup phase. For the initial standardized construction period, This is the initial resource delivery dispersion parameter. Parameters for the initial integration and coordination of personnel. For beta functions, This represents the probability density function value of the construction period fluctuation in the mid-term intensive construction phase. For the actual construction period in the middle stage, The location parameters of the mid-term log-normal distribution. The scaling parameter of the mid-log-normal distribution. The cumulative distribution function of the standard normal distribution. This is the upper limit of the mid-term construction period cutoff interval. This is the lower limit of the mid-term construction period cutoff interval. The planned construction period for the mid-term construction phase, This represents the probability density function value for schedule fluctuations during the later, final stages of the project. For the actual construction period in the later stage, For the number of rework sessions, For the intensity parameters of the later Poisson distribution, For the first Average construction period after each rework For the first Standard deviation of the project timeline after rework This is an approximation of the probability density function for the project duration fluctuation in the system's phase transition zone. For the derivation For the characteristic function, For the actual construction period of the transition zone, For the stability parameters of the Lévy distribution, The skewness parameter of the Lévy distribution. Let Lévy's distribution scale parameter be denoted as . The independent variable of the characteristic function is... For symbolic functions, For gamma function, For the first Daily project schedule delay For the first Daily schedule progress For the first Daily actual progress For the first Daily cumulative uncompensated progress deviation For the summation index, For the first Daily adjusted resource input As the baseline resource input, For resource adjustment coefficients, To compensate for the apportionment of subsequent days, This represents the number of days between the current date and the date of the delayed occurrence.

2. The method for optimizing engineering material procurement considering weather and price fluctuations according to claim 1, characterized in that, S2 specifically includes: S201. Use STL decomposition to perform time series decomposition on the material price series in the historical procurement data of engineering materials to obtain the decomposed time series. S202. Analyze the decomposed time series using the autocorrelation function, partial autocorrelation function, and unit root test to obtain the stationarity test results of the series. S203. Based on the stationarity test results of the sequence, establish a future price prediction model for commodities based on the ARIMA model to obtain the predicted commodity prices. S204. Based on the commodity price series and predicted commodity prices, construct a nonlinear price delivery function to obtain the price delivery balance relationship of different commodities from different suppliers.

3. The method for optimizing engineering material procurement considering weather and price fluctuations according to claim 1, characterized in that, The expression for the nonlinear price delivery function is as follows: in, For the first The first predicted path The price of goods at any given time For the predicted path number, In terms of time dimension, For the first Monthly forecast of commodity prices, For the first The price fluctuation disturbance term at time t under the path, Let the disturbance term follow a normal distribution. This is the fluctuation adjustment factor. The standard deviation of the fitted residuals of the nonlinear price delivery function. This represents the upper limit of the price prediction at time t with a 95% confidence level. This represents the lower bound of the price prediction at time t with a 95% confidence level. , and All are fusion coefficients. For the first Monthly forecast of commodity prices, This refers to the latest actual transaction price in the historical price series. This is the arithmetic average of all historical purchase prices.

4. The method for optimizing engineering material procurement considering weather and price fluctuations according to claim 1, characterized in that, The expression for the fluctuation data of material demand is as follows: in, For the first Daily final demand for supplies, i.e., data on fluctuations in supply demand. For the first Daily demand for basic supplies For the first Daily material demand compensation amount Let be the random fluctuation factor of demand on day t. For the first Daily demand for raw materials For the first Daily weather impact coefficient, For the first Daily rain resistance coefficient, For the first Daily rainfall, For the first Daily drag coefficient, For the first Daily average wind speed For the first Maximum daily gust wind speed.

5. The method for optimizing engineering material procurement considering weather and price fluctuations according to claim 1, characterized in that, The expression for the multi-objective optimization function in S4 is as follows: in, The total cost objective is optimized for multiple objectives. For material category indexing, This represents the total number of categories of materials required for the project. In terms of time dimension, To optimize the total number of days in the cycle, For the first Category of materials Daily purchase volume For the first Category of materials Daily purchase unit price, For the first The unit daily storage cost of this type of material, For the first Category of materials Daily inventory For the first Fixed cost per order for this type of material. For the first Category of materials Daily purchasing decision variables The material shortage risk objective is optimized for multiple objectives. For the first Category of materials Actual daily demand Safety stock coverage target optimized for multiple objectives. For the first Category of materials Daily safety stock level Heuristic inventory risk objectives for multi-objective optimization For the first The aging risk coefficient of high-risk materials This is the inventory aging and decay coefficient. For the first Category of materials Daily inventory backlog days.

6. The method for optimizing engineering material procurement considering weather and price fluctuations according to claim 5, characterized in that, The multi-objective optimization function is configured with constraints on storage capacity, procurement lead time, minimum order quantity, and safety stock lower limit. The expression for the storage capacity constraint is as follows: in, For the first Total daily storage occupancy of all materials For material category indexing, In terms of time dimension, For the first The percentage of each type of material by unit volume or unit weight. Total storage capacity To define the scope of application; The expression for the procurement lead time constraint is as follows: in, For the first Category of materials Daily inventory For the first Category of materials Daily inventory For the first Category of materials Daily purchase volume For the time dimension of purchasing orders, it means The supplies ordered today are Same-day delivery For the first Lead time for procurement of such materials For the first Category of materials Actual daily demand To define the scope of application; The expression for the minimum order quantity constraint is as follows: in, For the first Category of materials Daily purchase volume This is the upper limit of the maximum order quantity for supplies. For the first Minimum order quantity for this type of goods. For the first Category of materials Daily purchasing decision variables For the first The upper limit of the maximum order quantity for this type of material; The expression for the safety stock lower limit constraint is as follows: in, For the first Category of materials Daily safety stock level Safety stock factor For the first Category of materials Actual daily demand To constrain the scope of application.

7. A system for optimizing the procurement of engineering materials considering weather and price fluctuations, used to execute the method for optimizing the procurement of engineering materials considering weather and price fluctuations as described in any one of claims 1-6, characterized in that, include: The data acquisition and preprocessing module is used to acquire historical procurement data of engineering materials, supplier quotation data, weather data, project progress data and inventory data, and to perform data cleaning and standardization to obtain a standard dataset. The supply characteristic analysis module is used to analyze the characteristics of project schedule fluctuations and assess the fluctuations in material demand based on weather data, project progress data and inventory data in the standard dataset, and obtain the price trade-off relationship of different materials from different suppliers. The demand fluctuation assessment module is used to obtain project schedule fluctuation data and material demand fluctuation data by analyzing the characteristics of project schedule fluctuation and assessing the fluctuation of material demand based on weather data, project progress data and inventory data in the standard dataset. The multi-objective optimization solution module is used to construct a multi-objective optimization function based on the trade-off relationship between different materials and different suppliers, the fluctuation data of construction period and the fluctuation data of material demand. The module then uses the NSGA-II algorithm with a rolling window mechanism to solve the multi-objective optimization function and output the Pareto optimal solution set. The decision generation and visualization module is used to generate optimal decisions for engineering material procurement based on the Pareto optimal solution set using visualization technology.

Images