Product management method based on one-stop inspection platform

Abstract

The present invention relates to the technical field of logistics and supply chain management, and discloses a product management method based on a one-stop inspection platform, comprising the following steps: S1: constructing an optimal transportation problem model, defining the resource flow relationship between suppliers and orders, and calculating the optimal solution for suppliers to allocate to orders with the goal of minimizing transportation costs; S2: based on a game theory model, considering the strategic interactions between multiple suppliers, optimizing the supplier resource allocation strategy through game theory, and ensuring the balance of resource scheduling; S3: adopting a Markov decision process, adjusting the order flow and resource scheduling strategy through real-time feedback to adapt to the dynamic changes in business needs; S4: performing multi-objective optimization. The present invention realizes efficient resource scheduling and order flow through a dynamic optimization and real-time data feedback mechanism based on a one-stop inspection platform, ensuring maximum resource utilization, improving delivery timeliness, and optimizing costs and service quality.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics and supply chain management, and specifically to a product management method based on a one-stop inspection platform. Background Art

[0002] In modern society, with the rapid development of e-commerce and the logistics industry, efficient order processing and resource scheduling have become core to the competitiveness of major companies. To meet growing market demand, companies need to maximize the use of existing resources and reduce operating costs while ensuring service quality and delivery timeliness. In this context, improving the efficiency of the entire supply chain through reasonable resource scheduling and optimization strategies has become a pressing issue.

[0003] Existing technologies for resource scheduling and order management typically rely on static models and pre-defined rules. These approaches primarily optimize time and costs by pre-planning resource allocation and order flow to ensure efficient supply chain operations. These technologies can help companies reduce resource waste, improve the timeliness of order delivery, and ensure efficient resource utilization throughout operations. Most existing systems are well-suited to handling standardized resource allocation tasks, optimizing order allocation, and reducing the need for manual intervention, thereby improving efficiency.

[0004] However, there are still some shortcomings in the existing technology. In actual applications, market demand changes rapidly and supplier resources often fluctuate. The existing scheduling system is often unable to adapt to such changes, resulting in inflexible resource allocation and delayed order flow. In traditional solutions, due to the lack of a real-time feedback mechanism, the system is difficult to adjust resource allocation in a timely manner, resulting in uneven burdens on suppliers, some suppliers are overloaded with resources, while others are idle. In addition, many traditional systems still focus on the optimization of a single goal, such as cost control or time management, and fail to fully consider the needs of multi-objective balance. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the present invention provides a product management method based on a one-stop inspection platform, which solves the problems of inflexible resource scheduling, delayed order flow and lack of real-time data feedback adjustment in the existing technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A product management method based on a one-stop inspection platform includes the following steps:

[0007] S1: Construct an optimal transportation problem model, define the resource flow relationship between suppliers and orders, and calculate the optimal solution for suppliers to allocate to orders with the goal of minimizing transportation costs;

[0008] S2: Based on the game theory model, considering the strategic interactions among multiple suppliers, the supplier resource allocation strategy is optimized through game theory to ensure the balance of resource scheduling;

[0009] S3: Using a Markov decision process, we adjust order flow and resource scheduling strategies through real-time feedback to adapt to dynamic changes in business needs.

[0010] S4: Perform multi-objective optimization, comprehensively considering cost, time, and service quality, and calculate the optimal resource allocation plan to ensure optimal resource scheduling in a changing market environment;

[0011] S5: Dynamically adjust resource scheduling strategies based on real-time data to ensure optimal order flow efficiency and resource utilization.

[0012] Preferably, the constructing of the optimal transportation problem model includes:

[0013] Define the resource flow relationship between suppliers and orders, and establish a transportation cost matrix by considering order delivery time and supplier service capabilities;

[0014] Through the optimal transportation problem model, the linear programming method is used to calculate the optimal supplier and order matching plan to minimize the total transportation cost.

[0015] Preferably, the game theory-based model includes:

[0016] Set each supplier's strategy as its resource allocation method and analyze the competition and cooperation relationship among multiple suppliers;

[0017] Use the Nash equilibrium model to evaluate the optimal strategy of each supplier in the resource allocation game;

[0018] The Q-learning algorithm is used to dynamically optimize supplier strategies to ensure equilibrium during the game and achieve the optimal resource allocation solution.

[0019] Preferably, the Markov decision process includes:

[0020] Define the state space of order flow, including order processing progress, supplier load and resource utilization;

[0021] Defining the space of actions the platform can take, including decisions about selecting suppliers and allocating resources;

[0022] The optimal resource scheduling strategy is solved through the Bellman equation in the Markov decision process to achieve dynamic resource adjustment and ensure optimized order flow when business needs change.

[0023] Preferably, the multi-objective optimization includes:

[0024] Taking cost, time and service quality as objective functions, multi-objective optimization method is used to optimize them simultaneously;

[0025] Use non-dominated sorting genetic algorithm to solve the Pareto optimal solution between multiple optimization objectives to ensure the optimal balance between different objectives;

[0026] Dynamically adjust target weights based on real-time orders and supplier capabilities to ensure optimized resource scheduling in a volatile market environment.

[0027] Preferably, the dynamic adjustment of resource scheduling strategy includes:

[0028] Analyze current resource scheduling status based on real-time order volume, supplier load, and resource usage data;

[0029] The resource scheduling strategy is adjusted through adaptive algorithms to ensure that order processing time is minimized while ensuring resource utilization; the order flow path is dynamically optimized, and the real-time feedback mechanism enables the platform to adapt to demand fluctuations and adjust resource allocation in a timely manner.

[0030] Preferably, the steps of the Q-learning algorithm include:

[0031] Define the state space and action space. The state space includes supplier load, order progress, and resource status, while the action space includes supplier selection and resource allocation.

[0032] A reward function is set up to evaluate the benefits of each supplier adopting a certain strategy through real-time feedback, and the strategy is adjusted based on the reward; the Q value is continuously updated through the Q-learning algorithm, allowing suppliers to adjust resource allocation strategies according to current market conditions, thereby achieving the optimal game equilibrium.

[0033] Preferably, the Bellman equation includes:

[0034] According to the state space and action space of the platform, calculate the transition probability P(s′|s,a) of taking action in each state;

[0035] Evaluate the benefit of taking an action in each state based on the immediate reward R(s,a);

[0036] Solve the Bellman equation to obtain the optimal state value V(s), and optimize resource scheduling by selecting the best action.

[0037] The present invention provides a product management method based on a one-stop inspection platform. It has the following beneficial effects:

[0038] 1. This invention uses real-time data feedback and a dynamic optimization algorithm to enable the system to rapidly respond to market demand fluctuations. Compared to the fixed scheduling methods used in existing technologies, this invention ensures efficient and flexible order flow and resource allocation. Especially in situations of unstable market demand, the system can automatically adjust the optimization plan, avoiding order processing delays caused by fluctuating demand.

[0039] 2. This invention innovatively utilizes a multi-objective optimization algorithm to comprehensively consider multiple objectives, including cost, delivery time, and service quality, avoiding the overly single-objective optimization of traditional solutions. By dynamically adjusting weights, this invention minimizes costs and ensures on-time and high-quality order delivery without sacrificing other objectives.

[0040] 3. By dynamically adjusting resource allocation and optimizing goals, the platform not only reduces overall operating costs but also ensures efficient order flow. Compared to the static cost control used in existing technologies, this invention achieves dual optimization of cost and time, making the platform more precise in resource utilization and effectively avoiding resource waste and ineffective operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0042] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the specification of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0043] Please see the attached Figure 1 ,An embodiment of the present invention provides a product management method based on a one-stop inspection platform, comprising the following steps: S1: constructing an optimal transportation problem model, defining the resource flow relationship between suppliers and orders, and calculating the optimal solution for suppliers to allocate to orders with the goal of minimizing transportation costs;

[0044] S1's goal is to optimize the matching between suppliers and orders by minimizing transportation costs, thereby improving the efficiency of overall resource allocation.

[0045] The flow of resources between suppliers and orders is crucial on the platform. To ensure the platform can efficiently and accurately manage supplier resources, the Optimal Transportation Problem (OTP) model is used to optimize resource scheduling. This step defines a transportation cost matrix and calculates and determines the optimal match between each supplier and order to reduce costs and improve resource efficiency. The platform's resource scheduling not only considers cost but also factors such as service quality and delivery time, further ensuring the system can flexibly respond to diverse business needs.

[0046] In this embodiment, the purpose of constructing the optimal transportation problem model is to calculate the optimal pairing solution between suppliers and orders to ensure that transportation costs are minimized. First, it is necessary to describe the resource flow relationship by establishing a transportation cost matrix between suppliers and orders.

[0047] In the optimal transportation problem model, the platform has m suppliers and n orders. The transportation cost between suppliers and orders can be calculated based on a variety of factors, usually including the order delivery time, supplier service capabilities, transportation costs, etc. Assume C ij represents the service cost of supplier i for order j.

[0048] In this matrix, C ij It does not just represent a fixed cost, but is dynamically adjusted based on real-time business data (such as the urgency of order requirements, supplier service quality, etc.). For example, when an order has a strong time requirement, C ij The value of will increase accordingly, reflecting the additional cost required to meet the order time requirements. Similarly, the supplier's historical service quality can also affect C ij If the service quality provided by a supplier is poor, the corresponding element in the cost matrix will also increase, reducing the order allocation to that supplier.

[0049] The goal of the optimal transportation problem is to minimize the total transportation cost between all suppliers and orders. The objective function is as follows:

[0050]

[0051] Where: C ij X represents the transportation cost of supplier i providing services for order j, including time, cost, and service quality; ij is a decision variable. If order j is assigned to supplier i, then X ij =1, otherwise X ij =0; m is the number of suppliers; n is the order quantity.

[0052] To ensure that each order is processed by a supplier and that each supplier can only process one order, add the following constraints:

[0053] Each supplier can only process one order:

[0054]

[0055] This means that for each supplier i, there is only one order j assigned to that supplier.

[0056] Each order can only be assigned to one supplier:

[0057]

[0058] This means that each order j can only be processed by one supplier i.

[0059] By solving this optimization problem, the system can derive the optimal resource allocation plan and determine the orders allocated to each supplier.

[0060] To solve this optimal transportation problem, linear programming (LP) methods, such as the simplex method, or other optimization algorithms can be used. These algorithms can efficiently calculate the optimal solution required to minimize the objective function while satisfying the constraints. Through optimization calculations, the platform can accurately allocate resources, reduce costs, and improve efficiency.

[0061] In the actual business process, order requirements and supplier status often change. Therefore, in order to cope with these changes, the platform will dynamically adjust the transportation cost matrix. For example, when an order has a temporary change (such as an increase in demand or urgent delivery), the platform will update the C ij The optimal transportation plan is recalculated based on the value of .

[0062] These adjustments are based on real-time data from the platform, including suppliers' current load, resource usage, and order urgency. By dynamically adjusting the transportation cost matrix, the platform can better respond to market demand, ensure that resources are allocated appropriately, and avoid waste or overload.

[0063] Assume there are 3 suppliers and 5 orders. The system will construct a 5x3 cost matrix based on each supplier's processing capacity, service quality and other information, where each element C ij Represents the processing cost of supplier i for order j. For example:

[0064] Where: C 11 =10 means the cost for supplier 1 to process order 1 is 10; C 12 =15 means the cost for supplier 1 to process order 2 is 15; C 13=20 means the cost for supplier 1 to process order 3 is 20; C 21 =12 means the cost for supplier 2 to process order 1 is 12; C 22 =18 means the cost for supplier 2 to process order 2 is 18; C 23 =15 means the cost for supplier 2 to process order 3 is 15; C 31 =11 means the cost for supplier 3 to process order 1 is 11; C 32 =14 means the cost for supplier 3 to process order 2 is 14; C 33 =19 means the cost for supplier 3 to process order 3 is 19.

[0065] The platform solves this matrix using a linear programming algorithm and ultimately determines which supplier is best suited to assign each order to ensure that the total transportation cost is minimized.

[0066] This optimal transportation problem model is not limited to basic cost minimization but can be extended to incorporate more constraints and optimization objectives. For example, factors such as transportation time and service quality can be incorporated as additional optimization objectives. Furthermore, when supplier resources are limited, supplier capacity can be incorporated as a constraint to ensure the system fully utilizes supplier resources while avoiding overload.

[0067] In addition, in some embodiments, the platform can also make predictions based on historical order data, market demand trends, and other factors, and calculate and adjust the transportation cost matrix in advance to improve the system's response speed.

[0068] By defining a transportation cost matrix and applying a linear programming optimization algorithm, the platform can effectively calculate the optimal supplier-order matching solution. Dynamic adjustments and real-time updates ensure the system can adapt to changing market demands, improving resource utilization efficiency and order processing capabilities.

[0069] S2: Based on the game theory model, considering the strategic interactions among multiple suppliers, the supplier resource allocation strategy is optimized through game theory to ensure the balance of resource scheduling;

[0070] In a multi-supplier competitive environment, models that rely solely on cost minimization are insufficient to address strategic interactions between suppliers. Therefore, S2 employs game theory models, particularly the Nash equilibrium model, to optimize supplier resource allocation and ensure balanced resource scheduling across multiple suppliers. This process not only optimizes resource competition among suppliers but also achieves long-term stability and balance in supplier strategies.

[0071] In this embodiment, the core concept of the game theory model is to optimize each supplier's resource allocation decisions by simulating competition and cooperation among multiple suppliers, ultimately achieving a balanced resource allocation state through the game process. Specifically, we apply the concept of Nash equilibrium from game theory to calculate the supplier's optimal strategy, ensuring that, given the strategies of other suppliers, no supplier can achieve higher returns by unilaterally adjusting its own strategy.

[0072] Typically, game theory models involve multiple participants (in this case, suppliers), each of whom influences the system's resource scheduling and costs by selecting a resource allocation strategy. In this scenario, the supplier's goal is to select a resource allocation strategy that maximizes their gains in the game with other suppliers. By analyzing the interactions between suppliers, game theory can help the platform find an equilibrium resource allocation solution among them.

[0073] In the specific game theory model, the revenue function R of each supplier is i All depend on the strategy x chosen by supplier i i and other vendor selected strategies x j The ultimate goal of the game is to find a strategy combination that, when all suppliers adopt this strategy, no single supplier can achieve greater benefits by unilaterally changing its strategy. This is the definition of a Nash equilibrium.

[0074] Specifically, in this embodiment, it is assumed that m suppliers participate in the game, and each supplier i chooses a resource allocation strategy x i (For example, selecting the number of services, resource allocation, etc.) Each supplier’s revenue function R i Depends on the strategy x it chooses i and strategies of other suppliers x j The Nash equilibrium model in game theory requires finding a strategy combination (x1, x2, ..., x m ), so that each supplier meets the following conditions:

[0075]

[0076] in: Supplier i is taking the strategy The income when is the optimal resource allocation strategy of supplier i under Nash equilibrium; R i (x1,x2,...,x i ,...,x m ) is supplier i adopting other strategies x iThe benefits when x1, x2, ..., x m is the strategy combination of all suppliers; x1 represents the resource allocation strategy selected by supplier i.

[0077] This formula means that when the strategies of other suppliers are known, supplier i cannot obtain higher returns by unilaterally changing its own strategy, that is, the strategy combination is stable.

[0078] To effectively achieve dynamic optimization of the game theory model, this embodiment introduces the Q-learning algorithm, enabling suppliers to autonomously learn and adjust their resource allocation strategies based on market feedback. The Q-learning algorithm learns through interaction with the environment and continuously updates the Q value to optimize the supplier's decision-making strategy.

[0079] In Q-learning, the state space S represents all possible situations faced by the supplier, and the action space A represents all resource allocation strategies that the supplier can choose. By setting the reward function r i , the platform can evaluate the effect of a supplier's strategy and use the results to adjust future decisions. The update formula of Q-learning is:

[0080]

[0081] Where: Q(x i ,a i ) is supplier i in state x i Next take action a i Q value; α is the learning rate, which controls the speed of Q value update; r i Supplier i is taking action a i The immediate reward obtained after γ is the discount factor, which indicates the importance of future rewards. is in state x i ′, the Q value of the optimal strategy selected by supplier i.

[0082] This formula indicates that through interaction with the environment, the supplier updates its Q value based on the feedback of the current strategy, so that future decisions can continue to move closer to the optimal strategy.

[0083] In actual operations, suppliers' strategies are not fixed. As market demand, suppliers' resource load, and order urgency fluctuate, the platform needs to collect and analyze new data in real time to adjust its strategies. In one possible implementation, the platform collects real-time order data and supplier feedback to continuously adjust the input parameters of the game theory model, such as supplier processing capacity and market demand changes. This real-time data is used to dynamically update the cost matrix and reward function to optimize the supplier's strategy.

[0084] By leveraging the Nash equilibrium model in game theory, the platform optimizes resource allocation strategies across multiple suppliers, ensuring balanced resource scheduling and a balance between cooperation and competition among suppliers. Combined with the Q-learning algorithm, the platform continuously adjusts supplier resource allocation strategies based on real-time feedback, thereby optimizing overall scheduling.

[0085] S3: Using a Markov decision process, we adjust order flow and resource scheduling strategies through real-time feedback to adapt to dynamic changes in business needs.

[0086] S3 introduces the Markov Decision Process (MDP) and continuously adjusts resource scheduling and order flow strategies through a real-time feedback mechanism to ensure that the platform can adapt to changing market demands.

[0087] In this embodiment, the core of step S3 is to use a Markov Decision Process (MDP) model to dynamically adjust order flow and resource scheduling based on real-time feedback to achieve optimal resource allocation and processing results. This step is closely linked to the optimal resource allocation and game optimization results of steps S1 and S2, further enhancing the platform's flexibility and optimization capabilities.

[0088] Generally speaking, the Markov Decision Process (MDP) is an optimization decision model widely used in dynamic environments to select the optimal action based on the current state to maximize long-term goals. During the platform's resource scheduling and order flow, the system's state constantly changes over time and due to external factors (such as changes in orders and fluctuations in supplier load). Therefore, the platform relies on the MDP model to make optimal scheduling decisions at every moment to ensure efficient resource utilization and smooth order fulfillment.

[0089] Specifically, the MDP model consists of the following basic components:

[0090] The state space S represents all possible states of the platform system. In this embodiment, the states may include but are not limited to the processing status of the order, the resource load of the supplier, the change of market demand, etc.

[0091] The action space A represents all possible actions that the platform can take in each state. For example, the platform can choose different resource allocation schemes and scheduling strategies based on the current state.

[0092] The transition probability P(s′|s,a) represents the probability of transitioning from state s to a new state s′ after executing action a. This transition describes the evolution of the system state after each decision.

[0093] The reward function R(s,a) represents the immediate reward for taking action a in a given state s. The reward function is usually designed based on factors such as order processing efficiency, service quality, and cost-effectiveness.

[0094] The discount factor γ indicates the relative importance of future rewards. The larger the γ value, the more the platform values ​​long-term returns; the smaller the γ value, the more the platform focuses on current immediate rewards.

[0095] In the platform, the goal of the MDP model is to maximize the total reward of the system in the long term by making dynamic decisions about order flow and resource scheduling. The optimization goal of the platform can be expressed as a value function V(s):

[0096] V(s)=max a∈A [R(s,a)+γ·∑ s′ P(s′|s,a)·V(s′)];

[0097] Where: V(s) is the value function of state s, which represents the maximum expected reward obtained by executing the optimal policy starting from state s; R(s,a) is the immediate reward obtained after taking action a in state s; γ is a discount factor that measures the relative importance of future rewards; P(s′|s,a) is the probability of transitioning from state s to state s′ after taking action a; ∑ s′ P(s′|s,a)·V(s′) is the expected reward for taking action a in state s, resulting in all possible states s′. s is the current state of the platform. This state can include supplier load, order urgency, and service quality requirements. a is an action the platform can take, typically representing a resource allocation decision, such as selecting a supplier to assign an order to or adjusting resource scheduling.

[0098] This formula calculates the expected maximum reward for adopting different strategies (resource scheduling, order allocation, etc.) under the current state and selects the optimal strategy. This enables the platform to dynamically adjust resource allocation and order flow decisions, thereby maintaining efficient operations in a constantly changing environment.

[0099] As the platform's operations progress, the external market environment and internal operating conditions may change. Therefore, the MDP model in step S3 can be dynamically adjusted based on real-time feedback. For example, if a supplier's resource load becomes excessive, the platform incorporates this change into the MDP model through real-time feedback and adjusts the resource scheduling strategy to ensure a more balanced allocation of resources across suppliers and avoid overload.

[0100] In practice, the platform continuously monitors order progress, supplier resource status, and market demand, dynamically updating parameters such as the reward function R(s,a) and the transition probability P(s′|s,a). This enables the system to quickly adjust to market changes and maximize overall benefits.

[0101] To further enhance the optimization effect of the MDP model, the platform can incorporate Q-learning algorithms for reinforcement learning. Q-learning can optimize the platform's decision-making process through interaction with the environment, thereby achieving more refined resource scheduling in dynamic and complex environments.

[0102] Through Q-learning, the platform can continuously update the Q value and optimize decisions based on real-time feedback, so that each supplier's strategy gradually tends to be optimal, thereby maintaining efficient resource allocation and order flow in a complex and dynamic market environment.

[0103] By optimizing the MDP model, the platform can make optimal resource allocation decisions at any time, ensuring the system maintains efficient operation in a dynamic market environment. Combined with reinforcement learning using the Q-learning algorithm, the platform continuously optimizes supplier resource scheduling and order flow strategies, ensuring optimal resource allocation and service quality in the face of complex market fluctuations.

[0104] S4: Perform multi-objective optimization, comprehensively considering cost, time, and service quality, and calculate the optimal resource allocation plan to ensure optimal resource scheduling in a changing market environment;

[0105] S4 introduces a multi-objective optimization model. This model optimizes multiple objectives through a weighted combination approach to ensure the platform achieves an optimal resource allocation solution while taking into account cost, time, and service quality. This multi-objective optimization method is described in detail below, along with complete formulas and parameter definitions to ensure technical personnel clearly understand the meaning of each parameter.

[0106] In multi-objective optimization, the platform needs to optimize multiple objective functions simultaneously, each of which corresponds to an optimization goal of the platform. In this embodiment, the platform's goals include:

[0107] Minimizing the cost C(x) is to reduce the costs associated with resource scheduling, transportation, and supplier services.

[0108] Minimize the time T(x) to reduce the processing and delivery time of the order and try to improve the timeliness of the order.

[0109] Maximizing service quality Q(x) is to ensure order accuracy and customer satisfaction.

[0110] Usually, there may be conflicts between multiple objectives. Therefore, it is necessary to combine multiple objective functions into a comprehensive objective function through weighted summation. The objective function of the platform can be expressed as:

[0111] min(α·C(x)+β·T(x)-γ2·Q(x));

[0112] Where: C(x) represents the cost function, which measures the overall cost of resource scheduling and order flow; T(x) represents the time function, which measures the total time required from order placement to delivery; Q(x) is the quality function, which measures service quality or customer satisfaction; α, β, and γ2 are weight coefficients, α determines the weight of cost, β determines the weight of time, and γ2 determines the weight of quality; x is the decision variable, which represents the resource scheduling plan, order allocation strategy, etc.

[0113] Multi-objective optimization is not just about solving the objective function; it also requires considering the constraints in actual operations. These constraints may come from resource limitations, order requirements, supplier capabilities, etc. The following are some of the constraints that the platform may encounter:

[0114] Supplier resource constraints, which represent the maximum number of orders each supplier can handle;

[0115] The latest delivery time of the order to ensure that the order is delivered on time;

[0116] Service quality standards to ensure delivery meets customer requirements;

[0117] The system's maximum processing capacity avoids overload operation.

[0118] These constraints can be expressed using the following inequalities:

[0119]

[0120] Where: x ij is a decision variable indicating whether order j is served by supplier i (usually 0 or 1); S i is the maximum processing capacity of supplier i, indicating the maximum number of orders that each supplier can process; D j is the demand for order j, indicating the amount of resources or services required for the order; Tj is the delivery time of order j, indicating the delivery time point of the order; T max It is the maximum acceptable delivery time of an order, indicating that the order must be delivered before a certain time point.

[0121] These constraints ensure that the platform can meet the actual operational needs when performing multi-objective optimization and avoid unrealistic optimization solutions.

[0122] Multi-objective optimization problems are usually nonlinear, so the solution method needs to consider the balance between multiple objectives. The platform can choose from the following common solution methods:

[0123] Weighted approach: Multiple objective functions are combined into a single objective function through weighted summation. The platform achieves a balance between objectives by adjusting the weights of different objective functions.

[0124] ε-constraint method: Optimize one objective function as the main objective and transform the other objective functions into constraints. This method is suitable for situations where the objectives have different importance.

[0125] Pareto optimality: Find a set of solutions such that no single solution outperforms all others on all objectives. This method is suitable for situations where there is no clear priority between multiple objectives and the optimal solution can be determined by comparing the advantages and disadvantages of multiple solutions.

[0126] In some embodiments, the platform uses a weighted summation method and dynamically adjusts the weight coefficients α, β, and γ2, allowing the system to appropriately adjust the optimization target based on factors such as market demand and customer feedback. This enables the platform to obtain flexible optimization results in different business environments.

[0127] The platform uses a multi-objective optimization algorithm to generate a set of optimal resource scheduling solutions. To assess the effectiveness of these optimization results, the platform tracks and evaluates actual execution. For example, the platform monitors metrics such as actual order delivery time, actual costs, and customer feedback. If an optimization solution fails to achieve the desired results, the platform will use a feedback mechanism to readjust the objective function or weighting coefficients.

[0128] For example, in one embodiment, the platform processes multiple orders simultaneously and needs to allocate resources among multiple suppliers. In this process, the platform considers three key objectives:

[0129] Reduce transportation costs;

[0130] Ensure timely delivery of orders;

[0131] Improve customer satisfaction.

[0132] The platform uses a multi-objective optimization model to comprehensively consider the weight of each objective and arrive at a balanced resource scheduling solution. During actual execution, the platform dynamically adjusts the weight coefficients based on order changes to ensure that the optimization objectives are properly balanced under varying market demands.

[0133] By introducing multi-objective optimization models and constraints, the platform can maintain system operability and practical feasibility while considering the trade-offs between multiple objectives. Furthermore, the platform can dynamically adjust the optimization plan based on real-time feedback, thereby improving the system's flexibility and adaptability in complex environments.

[0134] S5: Dynamically adjust resource scheduling strategies based on real-time data to ensure optimal order flow efficiency and resource utilization. S5 introduces a dynamic adjustment mechanism based on real-time data, ensuring the platform can dynamically optimize order flow and resource scheduling strategies based on the latest real-time information. Through step S5, the platform can dynamically adjust the objective function, decision variables, and weight coefficients based on real-time feedback data in a changing market environment to optimize multiple objectives such as cost, delivery time, and service quality. This adjustment mechanism can respond to environmental changes in real time and ensure optimal resource scheduling.

[0135] Typically, the platform collects data related to orders, suppliers, and market demand in real time from multiple data sources. This data typically includes:

[0136] Supplier status data: including real-time data such as inventory levels, order processing capabilities, and transportation capabilities;

[0137] Order flow data: order receipt time, processing time, shipping status and delivery status, etc.

[0138] Market demand data: Real-time demand fluctuation data obtained by the platform based on factors such as customer orders, geographic location, and demand changes.

[0139] These data are collected in real time through IoT sensors, big data analysis tools, API interfaces, etc., and transmitted to the platform's real-time data processing system.

[0140] In this embodiment, the platform integrates big data analytics and real-time sensor monitoring to monitor real-time supplier and market dynamics. For example, by monitoring suppliers' inventory information, transportation capacity, and order processing capabilities in real time, the platform can dynamically adjust resource allocation strategies to ensure optimal resource scheduling.

[0141] In steps S1 through S4, the platform uses optimization models to calculate optimal resource scheduling solutions. However, these solutions may no longer be optimal in the face of a dynamically changing market and resource environment. Therefore, in step S5, the platform dynamically adjusts the aforementioned optimization models (e.g., optimal transportation problem models, game theory models, Markov decision process models, and multi-objective optimization algorithms) based on real-time data to optimize the current resource scheduling and order flow strategies.

[0142] In this embodiment, the core of dynamic adjustment is to update the objective function and constraints based on real-time data, and dynamically adjust the weight coefficient α t , β t , γ t and the decision variable X ij Through these adjustments, the platform can continue to optimize resource allocation in a real-time changing environment.

[0143] Specifically, the platform dynamically updates the cost function C(x), time function T(x) and quality function Q(x) through real-time data analysis, and adjusts the weight coefficient in the objective function to ensure efficient resource scheduling in a changing market environment.

[0144] In step S5, the platform dynamically adjusts the objective function in step S4 through real-time data feedback to reflect current market changes and actual needs.

[0145] min(α t C(x)+β t T(x)-γ t Q(x));

[0146] Among them: C(x) is the cost function, which measures the total cost of resource scheduling, usually including transportation costs, supplier costs, inventory management, etc. This function is dynamically updated based on real-time supplier load and market demand; T(x) is the time function, which measures the total time required from order placement to delivery, including transportation time, processing time, etc. In real-time adjustments, the system will update the time function based on the actual transportation progress and order status; Q(x) is the quality function, which measures service quality or customer satisfaction, including order accuracy, delivery timeliness, customer satisfaction, etc. The quality function will be dynamically adjusted based on customer feedback, order accuracy, etc. in real-time data; α t , β t , γ t is a weight coefficient that represents the relative importance of each objective in real-time optimization. During real-time adjustment, the platform adjusts these coefficients based on factors such as market demand, supplier status, and order urgency. For example, when market demand increases sharply, the platform may increase the time weight and reduce the cost weight; ijis a decision variable, indicating whether order j is served by supplier i. Based on real-time data and resource load, the platform adjusts the matching relationship between orders and suppliers in real time.

[0147] To ensure that the optimization plan meets the constraints of real-world operations, the platform dynamically updates the constraints based on real-time data. For example, when a supplier's resource load reaches its upper limit, the system adjusts the resource allocation strategy to ensure that the supplier's processing capacity is not exceeded, using the constraints expressed in the aforementioned inequality.

[0148] These constraints will be adjusted as real-time data is updated. For example, when the platform monitors the supplier's load in real time, if it finds that a supplier's capacity is close to saturation, the system will automatically update the supplier's maximum processing capacity S i , and adjust the order allocation strategy accordingly.

[0149] When implementing real-time adjustments, the platform requires an effective feedback mechanism to ensure the rationality and timeliness of dynamic adjustments. By continuously monitoring order status, customer feedback, market demand, and other factors, the platform can evaluate the effectiveness of current strategies in real time and adjust optimization strategies based on actual conditions.

[0150] For example, if a supplier's shipping capacity decreases, causing order delivery delays, the system can increase the weight of delivery time in real time and adjust resource scheduling strategies to prioritize orders to suppliers with more resources to ensure that delays are minimized.

[0151] For example, while processing an order, the platform discovers through real-time data that a supplier's processing capacity has reached capacity and that demand for orders has increased dramatically. Using the dynamic adjustment mechanism in step S5, the platform adjusts the objective function in real time, increasing the weighting of time and prioritizing orders to suppliers with greater inventory and shipping capacity, ensuring on-time delivery and minimizing costs.

[0152] Through real-time data feedback, the platform can continuously adjust optimization goals such as cost, time, and quality, enabling it to maintain efficient operation in a constantly changing environment.

[0153] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. The product management method based on the one-stop inspection platform is characterized by: The following steps are involved: S1: Construct an optimal transportation problem model, define the resource flow relationship between suppliers and orders, and calculate the optimal solution for suppliers to allocate to orders with the goal of minimizing transportation costs; S2: Based on the game theory model, considering the strategic interaction between at least two suppliers, the supplier resource allocation strategy is optimized through game theory to ensure the balance of resource scheduling; S3: Using a Markov decision process, we adjust order flow and resource scheduling strategies through real-time feedback to adapt to dynamic changes in business needs. S4: Perform multi-objective optimization, comprehensively considering cost, time, and service quality, and calculate the optimal resource allocation plan to ensure optimal resource scheduling in a changing market environment; S5: Dynamically adjust resource scheduling strategies based on real-time data to ensure optimal order flow efficiency and resource utilization.

2. The product management method based on the one-stop inspection platform according to claim 1 is characterized in that: The construction of the optimal transportation problem model includes: Define the resource flow relationship between suppliers and orders, and establish a transportation cost matrix by considering order delivery time and supplier service capabilities; Through the optimal transportation problem model, the linear programming method is used to calculate the optimal supplier and order matching plan to minimize the total transportation cost.

3. The product management method based on the one-stop inspection platform according to claim 1 is characterized in that: The game theory-based model includes: Set each supplier's strategy as its resource allocation method and analyze the competition and cooperation relationship between at least two suppliers; Use the Nash equilibrium model to evaluate the optimal strategy of each supplier in the resource allocation game; The Q-learning algorithm is used to dynamically optimize supplier strategies to ensure equilibrium during the game and achieve the optimal resource allocation solution.

4. The product management method based on the one-stop inspection platform according to claim 1 is characterized in that: The Markov decision process includes: Define the state space of order flow, including order processing progress, supplier load and resource utilization; Defining the space of actions the platform can take, including decisions about selecting suppliers and allocating resources; The optimal resource scheduling strategy is solved through the Bellman equation in the Markov decision process to achieve dynamic resource adjustment and ensure optimized order flow when business needs change.

5. The product management method based on the one-stop inspection platform according to claim 1 is characterized in that: The multi-objective optimization includes: Taking cost, time and service quality as objective functions, an optimization method that can optimize at least two objectives simultaneously is adopted; Using a non-dominated sorting genetic algorithm to solve the Pareto optimal solution between the at least two optimization objectives to ensure that an optimal balance is found between different objectives; Dynamically adjust target weights based on real-time orders and supplier capabilities to ensure optimized resource scheduling in a volatile market environment.

6. The product management method based on the one-stop inspection platform according to claim 1 is characterized in that: The dynamic adjustment of resource scheduling strategy includes: Analyze current resource scheduling status based on real-time order volume, supplier load, and resource usage data; Adjust resource scheduling strategies through adaptive algorithms to minimize order processing time while ensuring resource utilization; Dynamically optimize the order flow path, and the real-time feedback mechanism enables the platform to adapt to demand fluctuations and adjust resource allocation in a timely manner.

7. The product management method based on the one-stop inspection platform according to claim 3 is characterized in that: The steps of the Q-learning algorithm include: Define the state space and action space. The state space includes supplier load, order progress, and resource status, while the action space includes supplier selection and resource allocation. Set a reward function, evaluate the benefits of each supplier adopting a certain strategy through real-time feedback, and adjust the strategy based on the reward; By continuously updating the Q value through the Q-learning algorithm, suppliers can adjust their resource allocation strategies according to the current market conditions, thereby achieving the optimal game equilibrium.

8. The product management method based on the one-stop inspection platform according to claim 4 is characterized in that: The Bellman equation includes: According to the state space and action space of the platform, calculate the transition probability P(s′|s,a) of taking action in each state; Evaluate the benefit of taking an action in each state based on the immediate reward R(s,a); Solve the Bellman equation to obtain the optimal state value V(s), and optimize resource scheduling by selecting the best action.

Images