A production optimization method and system integrating logistics management

Abstract

The present invention discloses a production optimization method and system combined with logistics management, and relates to the field of production optimization technology. A production optimization system combined with logistics management includes: a personalized order parsing module, a logistics status acquisition module, an internal logistics modeling module, an external logistics optimization module, a supply strategy management module, a joint scheduling optimization module, a scheduling execution module, and a congestion monitoring and simulation feedback module. The present invention structurally expresses the external logistics resource supply strategy as a transportation time window matrix and a dynamic capacity allocation table, and introduces a joint scheduling optimization model as a boundary condition, so that the scheduling system can dynamically adjust the production and logistics coordination plan according to the external resource status, effectively enhancing the scheduling strategy's adaptability to external supply changes.

Description

Technical Field

[0001] This invention relates to the field of production optimization technology, and in particular to a production optimization method and system that integrates logistics management. Background Technology

[0002] With the increasing complexity of manufacturing systems and the growing personalization of order demands, production optimization is gradually shifting from static scheduling to collaborative optimization that dynamically adapts to resources and demand. Existing production scheduling methods are mostly based on order data and process flows, combining manufacturing resource capacity constraints to allocate tasks and prioritize them, providing relatively mature support for internal factory scheduling logic. In some high-end manufacturing scenarios, production process flow diagrams based on graph structure modeling are already available, enabling dynamic scheduling optimization at the process and equipment levels.

[0003] Meanwhile, factory logistics systems, as a crucial support connecting various production nodes, are gradually evolving towards automation and real-time operation. For internal factory logistics, some studies have proposed using Petri nets, queue networks, and other methods to model node channels and transportation resources to assess equipment occupancy and path reachability. For external logistics scheduling, existing transportation management systems or supply chain visualization platforms have achieved real-time coordination of vehicle scheduling and route planning, but these often exist as independent systems, relying primarily on pre-defined rules or manual intervention for interface connections with workshop scheduling systems.

[0004] In the manufacturing of customized products, order heterogeneity and real-time delivery pressures pose challenges to traditional scheduling systems. In manufacturing scenarios characterized by diverse product types, small batches, and sensitive delivery cycles, existing solutions still have room for improvement in their integration and coordination capabilities between logistics management and production scheduling. Particularly when external transportation resource supply strategies influence production scheduling decisions, there is a lack of a mechanism to dynamically incorporate these strategies into the production system's optimization process, creating a closed-loop collaboration across the modeling, scheduling, and feedback layers. This has become one of the core challenges in current integrated manufacturing logistics scheduling. Summary of the Invention

[0005] A production optimization method that integrates logistics management includes:

[0006] Receive personalized customer orders and extract order feature data;

[0007] Personalized customer orders for customized products are called customized orders. Before the actual customized order is fully confirmed, the customer's configuration intention is obtained and cluster analysis is performed to generate typical configuration clusters, which are used to construct corresponding shadow tasks. The shadow tasks are introduced into the joint scheduling optimization model as pre-scheduling inputs.

[0008] Real-time collection of status information of internal and external logistics resources in the factory, referred to as internal logistics information and external logistics information respectively, and modeling of internal logistics information to generate internal logistics capacity map of the factory;

[0009] The factory, logistics supplier and its subordinate external transportation fleet are modeled as intelligent agents. A dynamic game model based on a multi-agent reinforcement learning framework is used to generate the optimal response strategy for each intelligent agent by adopting a hierarchical game strategy, and output the external logistics resource supply strategy.

[0010] Based on order characteristic data, internal factory logistics capacity map, external logistics resource supply strategy and shadow tasks, a joint scheduling optimization model based on graph structure modeling is established to perform collaborative scheduling optimization of production tasks and internal logistics paths.

[0011] Based on the results of collaborative scheduling optimization, operation instructions and transportation scheduling tasks are generated for each production unit and logistics node.

[0012] During the execution of transportation scheduling tasks, internal logistics information is collected in real time to obtain congestion indicators for transportation scheduling tasks. When the congestion indicators are abnormal, the transportation scheduling tasks are dynamically adjusted.

[0013] When the congestion index of the transportation scheduling task is still abnormal after dynamic adjustment, the path conflict of the transportation scheduling task is simulated and predicted, and the simulation prediction results are fed back to the joint scheduling optimization model to trigger adaptive rescheduling and form a decision closed loop.

[0014] As a preferred embodiment of the present invention, the generation of the factory internal logistics capacity diagram includes:

[0015] A Petri net modeling tool is used to construct an internal logistics capability map of the factory, which includes the topological connections between logistics nodes and resource transfer rules. The internal logistics information includes the occupancy status and service capabilities of logistics nodes. Logistics nodes include fixed facilities, mobile devices and external interfaces within the factory. The resource transfer rules are constructed based on the service capabilities of each logistics node, including task processing cycle time, path accessibility status and reachability conditions.

[0016] As a preferred embodiment of the present invention, the method of generating the optimal response strategy for each agent using a hierarchical game strategy includes:

[0017] The factory, logistics supplier and its subordinate external transportation fleet are modeled as intelligent agents for reinforcement learning. Each intelligent agent constructs an observation state set based on order feature data and external logistics information, and pre-sets a set of corresponding optional actions based on its role definition.

[0018] A dynamic game model is constructed, in which the factory acts as the leader agent to generate the leader's strategy; the logistics supplier and its subordinate external transportation fleet act as follower agents, and after acquiring the leader's strategy, they use reinforcement learning to optimize their own response strategy. The optimization process is based on a set reward function, which includes a weighted combination of transportation cost, on-time delivery rate, and path load balancing.

[0019] The optimization results are output as an external logistics resource supply strategy, including transportation route allocation, vehicle scheduling, supply time window and task undertaking relationship.

[0020] As a preferred embodiment of the present invention, the external logistics resource supply strategy is expressed in a structured manner as a transportation time window matrix and a dynamic capacity allocation table, which serve as the input boundary conditions of the joint scheduling optimization model and are updated in real time through the intelligent agent communication interface.

[0021] As a preferred embodiment of the present invention, the joint scheduling optimization model includes constructing an order-driven production process flow diagram based on order feature data and an internal logistics network diagram based on the factory's internal logistics capacity diagram. The process flow diagram uses order work units and production equipment as nodes and process dependencies as edges. The internal logistics network diagram uses logistics nodes as nodes and topological connections as edges. The production equipment nodes in the production process flow diagram and the nearest logistics nodes in the internal logistics network diagram are automatically mapped to form a joint graph model. The objective function is set to minimize production task delays and maximize equipment and transportation resource utilization, while the resource transfer rules of the factory's internal logistics capacity diagram and external logistics resource supply strategies are used as boundary conditions for collaborative scheduling optimization.

[0022] When a customized order has not been fully confirmed, but there is a typical configuration cluster generated by configuration intent clustering, shadow tasks are introduced into the joint graph model as pre-scheduled tasks for resource pre-allocation and path pre-scheduling.

[0023] As a preferred embodiment of the present invention, the generation and scheduling of the shadow task includes:

[0024] Before a customized order is fully confirmed, the customer's configuration intent is constructed as a vector and clustered to generate representative typical configuration clusters. For each configuration cluster, a typical process flow and resource requirement template are generated, and shadow tasks are constructed accordingly. The shadow tasks are introduced into the joint scheduling optimization model in the form of soft constraints to achieve resource pre-allocation and path pre-scheduling before the customized order is confirmed. When the actual order is placed, the corresponding shadow task is matched and activated. Unmatched shadow tasks are released after a set time window to improve the adaptability to the diversity and uncertainty of customized orders.

[0025] The configuration intent includes customer configuration behavior data on the customization platform and historical order data.

[0026] As a preferred technical solution of the present invention, the method for determining the abnormality of the congestion index includes task congestion and local logistics capacity bottlenecks, and the congestion index is calculated in batches at fixed time intervals; the dynamic adjustment includes work order priority, path selection and batch division, and the dynamic adjustment must meet the constraints of the process dependency relationship of the production process flow chart.

[0027] As a preferred technical solution of the present invention, the method for determining task congestion is to obtain the occupancy status of each logistics node in the internal logistics capacity diagram of the factory, calculate the variance of the resource occupancy rate in the occupancy status, and determine that task congestion occurs when the variance of the resource occupancy rate exceeds a set threshold and the duration exceeds a set duration. The method for determining local logistics capacity bottlenecks is to identify logistics nodes in the internal logistics capacity diagram of the factory where resource request blocking occurs continuously, and detect whether the node has a situation where the task waiting time exceeds a set threshold in multiple consecutive scheduling cycles. If so, it is determined to be a local logistics capacity bottleneck.

[0028] A production optimization system integrating logistics management includes:

[0029] Personalized order parsing module: used to receive personalized customer orders and extract order feature data;

[0030] Custom Intent Clustering Module: Used to obtain customer configuration intents before customized orders are fully confirmed, cluster them, generate typical configuration clusters, build corresponding shadow tasks, and manage the introduction, activation, and release of shadow tasks;

[0031] Logistics Status Acquisition Module: Used to collect status information of internal and external logistics resources in the factory, and generate internal logistics information and external logistics information;

[0032] Internal logistics modeling module: used to construct a factory internal logistics capacity map based on internal logistics information, and to describe the topology and resource transfer rules of logistics nodes using Petri net modeling tools;

[0033] External logistics optimization module: This module models factories, logistics suppliers, and their subordinate transportation fleets as intelligent agents, constructs a dynamic game model, and generates external logistics resource supply strategies.

[0034] Supply strategy management module: used to express the supply strategy of external logistics resources in a structured way, and realize its timed update through the communication interface, and generate boundary conditions for the joint scheduling model to call;

[0035] The joint scheduling optimization module establishes a joint scheduling optimization model based on order characteristic data, internal factory logistics capacity map, external logistics resource supply strategy and shadow tasks, and performs collaborative scheduling optimization of production tasks and internal logistics paths.

[0036] The scheduling and execution module generates work instructions and transportation scheduling tasks for each production unit and logistics node based on the results of collaborative scheduling optimization.

[0037] The congestion monitoring and simulation feedback module is used to monitor whether the congestion indicators of transportation scheduling tasks are abnormal, and when the abnormality persists, it performs path conflict simulation prediction and feeds the simulation results back to the joint scheduling optimization module to trigger rescheduling operations.

[0038] The present invention has the following advantages:

[0039] This invention models factories, logistics suppliers, and their transportation fleets as intelligent agents using reinforcement learning, constructs a dynamic game model, and generates external logistics resource supply strategies, including transportation route allocation, vehicle scheduling, supply time windows, and task acceptance relationships. This solves the problems of static setting of external logistics scheduling, delayed response, and lack of intelligent optimization in traditional systems.

[0040] This invention, through configuration intent clustering and shadow task pre-scheduling mechanism, addresses typical manufacturing scenarios where customized product orders are highly uncertain, concentrated, and subject to sudden impacts. Before orders are confirmed, configuration trends are identified and representative shadow tasks are constructed to participate in resource scheduling and path planning in advance, effectively alleviating the problem of delayed scheduling response. This realizes a pre-scheduling strategy for customized orders, improving scheduling response speed and resource adaptability in a diversified and high-frequency customization environment.

[0041] This invention expresses the external logistics resource supply strategy in a structured manner as a transportation time window matrix and a dynamic capacity allocation table, and introduces them as boundary conditions into a joint scheduling optimization model. This enables the scheduling system to dynamically adjust the production and logistics coordination plan according to the status of external resources, effectively enhancing the adaptability of the scheduling strategy to changes in external supply.

[0042] This invention constructs a joint scheduling optimization model based on order feature data, factory internal logistics capacity diagram, and external supply strategy. It integrates production process flow diagram and internal logistics network diagram, and introduces dynamic external supply strategy and internal logistics status feedback mechanism to realize a fully intelligent closed loop from order input to scheduling output and feedback adjustment, so as to meet the real-time response requirements of highly personalized production.

[0043] This invention designs a congestion identification mechanism for the internal logistics capacity map of a base factory. By utilizing resource occupancy variance and task waiting time indicators, it dynamically identifies task congestion and local logistics capacity bottleneck nodes. When anomalies persist, it performs path conflict simulation prediction and feeds back to the scheduling model to form a scheduling closed loop, thereby improving the real-time response capability to local imbalances and path conflicts. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only schematic diagrams of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0045] Figure 1 This is a schematic diagram of a production optimization system that combines logistics management, as used in an embodiment of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0047] Example 1: A production optimization method combining logistics management, comprising the following steps:

[0048] Step S1: Receive personalized customer orders and extract order feature data, including but not limited to the order's process requirements, delivery deadline, material requirements and resource constraints;

[0049] Specifically, this includes:

[0050] Process requirements include, but are not limited to, information such as customer-specified product processing flow, key process sequence, and customized parameters (such as specific heat treatment curves and specific tool paths), which are parsed from the MES system or order input interface and used as the basis for constructing nodes and edges when building the production process flow diagram;

[0051] Delivery timeframes include, but are not limited to, expected delivery time and tolerable advance / delay tolerance time windows, which are derived from the order management system or customer-side configuration platform. This information is directly used in the joint scheduling model to set production task priorities and generate path scheduling timing constraints.

[0052] Material requirements include, but are not limited to, the codes, specifications, expected quantities, and supply cycle constraints of the raw materials required for the product. These are extracted from the material master data in the ERP system or explicitly set by the customer during the order acceptance process. This field is used in subsequent scheduling to ensure that the material arrival time matches the task time of the logistics node.

[0053] Resource constraints include, but are not limited to, customer-specified production lines, equipment preferences, exclusive processing resources (such as a certain model of CNC), and transportation mode preferences (such as restrictions on cold chain transportation or nighttime delivery). Such data can be set as optional parameters in the order configuration platform or can be derived from learning rules from historical orders to limit the candidate solution space of the joint scheduling model.

[0054] For example, a specific order feature data is as follows: An order for customized mechanical parts, with a process requirement specifying "heat treatment followed by grinding," a delivery deadline of "delivery before June 20th, allowing 1 day earlier or 2 days later," material requirements of "20 Φ100mm 45# steel bars," and resource constraints of "processing only on line 4, no nighttime shipments allowed." This order feature data will form an order feature vector, serving as the core input for downstream multi-agent modeling and scheduling optimization.

[0055] Step S2: For personalized customer orders with customized products, the order content is called a customized order. Before the actual customized order is fully confirmed, the customer's configuration intention is obtained and cluster analysis is performed to generate typical configuration clusters, which are used to construct corresponding shadow tasks. The shadow tasks are introduced into the joint scheduling optimization model as pre-scheduling inputs.

[0056] The generation and scheduling of the shadow tasks include:

[0057] Before a customized order is fully confirmed, the customer's configuration intent is constructed as a configuration vector. Cluster analysis is performed on multiple configuration vectors to obtain several representative typical configuration clusters. Typical process flow and resource requirement templates are generated for each configuration cluster, and shadow tasks are constructed accordingly. The shadow tasks are introduced into the joint scheduling optimization model in the form of soft constraints to achieve resource pre-allocation and path pre-scheduling before the customized order is confirmed. When the customized order is officially confirmed and matches a certain configuration cluster, the system automatically activates the corresponding shadow task and incorporates it into the formal scheduling process. If the shadow task is not matched to an actual order within a set time window, the system releases or clears the shadow task to free up system resources.

[0058] The configuration intent includes customer configuration behavior data and historical order data on the customization platform. The configuration behavior data includes the configuration parameter information selected or preferred by the customer on the customization platform and click behavior records. This mechanism significantly enhances the proactive response capability and scheduling flexibility in the face of the diversity and uncertainty of customized products by the structured expression of potential customization needs and the introduction of a joint scheduling optimization model.

[0059] The shadow task refers to a virtual scheduling task unit generated by configuration intent clustering before a customized order is formally confirmed, used to participate in the joint scheduling optimization model's optimization calculations in advance. Shadow tasks are not directly bound to actual materials and actions, but they have similar process paths and resource requirement descriptions to future orders, and are identified and processed by the joint scheduling optimization model.

[0060] The specific form of the shadow task includes the following key fields: Task ID, used to uniquely identify the shadow task; Source configuration cluster ID, identifying the configuration cluster number corresponding to the shadow task; Expected process flow, a typical process sequence generated by the configuration cluster, such as [rough processing → heat treatment → fine processing → inspection]; Expected resource requirements, including the required production equipment type, logistics node type, processing time, cycle time, etc.; Virtual time window, the predicted start and finish time interval, used for planning the occupied window; Execution priority weight, set to be lower than the formal order task, which can be used as a soft constraint in the scheduling objective function; Status flags, including three states: unmatched, activated, and released, used for lifecycle management.

[0061] Shadow tasks are added to the joint graph model as variable scheduling entities, participating in scheduling solutions together with formal order tasks, but they are not executed by the scheduling system and are only used to influence path planning and resource allocation strategies.

[0062] Example: A configuration cluster represents a frequently occurring combination of "body + premium interior + large wheels". The system generates a process path for it as "painting → interior assembly → final assembly → quality inspection", with resource requirements of equipment E2 / E5, AGV-2 channel, and buffer area C1. The time window is 14:00~17:00 tomorrow. This shadow task participates in scheduling with low weight in the joint scheduling model. Once the system detects that the customer has placed an order and confirmed the corresponding configuration, this task is automatically activated and becomes a formal scheduling item.

[0063] Step S3: Collect real-time status information of internal and external logistics resources in the factory, referred to as internal logistics information and external logistics information, respectively. Model the internal logistics information to generate a factory internal logistics capacity map.

[0064] The generated internal logistics capacity diagram of the factory includes:

[0065] A Petri net modeling tool is used to construct an internal logistics capability map of the factory, which includes the topological connections between logistics nodes and resource transfer rules. The internal logistics information includes the occupancy status and service capabilities of logistics nodes. Logistics nodes include fixed facilities, mobile devices and external interfaces within the factory. The resource transfer rules are constructed based on the service capabilities of each logistics node, including task processing cycle time, path accessibility status and reachability conditions.

[0066] The fixed facilities include transport channels and loading / unloading stations; the mobile equipment includes, but is not limited to, automated guided vehicles (AVGs), forklifts, manual transport vehicles and other in-plant transport vehicles; the external interface is the loading / unloading area that connects with the transport fleet.

[0067] The specific details of the internal logistics information are as follows:

[0068] The occupancy status of logistics nodes refers to the current status of each logistics unit (such as loading and unloading area, conveyor belt start and end points, buffer area, and transshipment platform), such as "idle", "in operation", "queuing", etc. It is obtained in real time through embedded sensors, RFID readers, MES interfaces, etc., and serves as the status mapping of each location (Place) in the Petri net.

[0069] The service capacity of a logistics node includes parameters such as the number of tasks it can handle per unit time (e.g., 6 transports per hour on an AGV line), processing unit cycle time (e.g., time taken for each loading), and maximum buffer capacity (e.g., 5 pallets that can be accommodated simultaneously on a conveyor belt). This data can be obtained from the process master data system, historical operating analysis system, or manually calibrated.

[0070] Topological connectivity refers to the accessibility and path connectivity structure between logistics nodes. For example, "from the raw material warehouse to the injection molding workshop, it is necessary to pass through channel A2 and elevator B1", which can be automatically generated by importing CAD layout diagrams or PLC configuration diagrams into modeling tools.

[0071] Resource transfer rules represent the conditions for task migration between different nodes, such as "the transfer from node X to node Y can only be triggered when the AGV is idle and the path is not occupied". This logic is implanted as a transition triggering rule in Petri net for dynamic simulation and state evolution.

[0072] Example: In a factory AGV system, node A is the loading station, node B is the processing area, and node C is the unloading station. If node B currently has a queue length of 2, a maximum processing capacity of 3, a processing time of 10 minutes per unit, and currently has 5 AGVs with a loading capacity of 1 trip per AGV, then the current effective service capacity of the A→B→C path can be obtained through modeling, and the path availability and task load can be evaluated accordingly. The internal logistics capacity map of the factory constructed by the above modeling results will serve as the basic structural input for subsequent path planning, scheduling optimization, and bottleneck identification.

[0073] Step S4: Model the factory, logistics supplier and its subordinate external transportation fleet as intelligent agents. Through a dynamic game model built on a multi-agent reinforcement learning framework, a hierarchical game strategy is adopted to generate the optimal response strategy for each intelligent agent and output the external logistics resource supply strategy.

[0074] The strategy for generating the optimal response for each agent using a hierarchical game strategy includes:

[0075] The factory, logistics supplier and its subordinate external transportation fleet are modeled as intelligent agents for reinforcement learning. Each intelligent agent constructs an observation state set based on order feature data and external logistics information, and pre-sets a set of corresponding optional actions based on its role definition.

[0076] A dynamic game model is constructed, in which the factory is used as the leader agent to generate the leader strategy; the logistics supplier and its subordinate external transportation fleet are used as follower agents. After acquiring the leader strategy, they use reinforcement learning to optimize their own response strategy. The optimization process is based on a set reward function and is calculated only based on the local state of their current transportation task; the reward function includes a weighted combination of transportation cost, on-time delivery rate, and path load balancing.

[0077] The optimization results are output as an external logistics resource supply strategy, including transportation route allocation, vehicle scheduling, supply time windows, and task acceptance relationships. The external logistics resource supply strategy is expressed in a structured form as a transportation time window matrix and a dynamic capacity allocation table, which serve as the input boundary conditions for the joint scheduling optimization model and are updated in real time through the agent communication interface.

[0078] The set of observed states comes from the order feature data in step S1 and the external logistics information in step S3 (such as vehicle idle status, task waiting queue length, and current road traffic load); for example, the observed state of transport fleet A is [fleet idle rate 60%, route X number of waiting tasks 3, next time period transport window opening time t=5min];

[0079] The set of optional actions is set according to the type of intelligent agent: For factories, optional actions include at least whether to split orders, set task release time, select task merging strategy, and adjust resource priority allocation method; for transportation fleets, optional actions include at least whether to accept the current order, select which route to transport, plan departure time, and participate in task reassignment; for logistics providers, optional actions include at least allocating the number of vehicles, specifying available service time windows, and adjusting the cargo loading strategy of each fleet.

[0080] The reward function design comprehensively considers objectives such as logistics costs (unit transportation costs (obtained from actual logistics operators), empty-run rate), on-time delivery rate (whether the task is completed within the specified window), and route balance (whether the same road segment is overloaded). For example, if the transportation fleet chooses route Y to complete the task 2 minutes earlier but incurs an additional 30 yuan in fuel costs, the strategy score is calculated based on the weighted settings (cost as 0.4, on-time rate as 0.6).

[0081] The leader-follower model is based on Stackelberg game logic, in which the factory, as the producer of production plans, imposes prior constraints on downstream agents on its production schedule and demand release time windows. Logistics suppliers and transportation fleets then respond by optimizing and allocating resources under these constraints. This hierarchical structure can ensure the overall consistency of the system.

[0082] The data structure of external logistics resource supply strategies includes, but is not limited to: transportation route allocation: such as "Fleet A is assigned to route P1, with an estimated route load of 80%"; vehicle scheduling: such as "Vehicle 3 is scheduled to depart at 15:00 and is expected to arrive at 18:20"; supply time window: such as "The available window for loading and unloading area B is 16:00~17:00"; task acceptance relationship: such as "Order O1 is accepted by logistics supplier L and assigned to fleet T1 for transportation".

[0083] The external logistics resource supply strategy is expressed in a structured form as follows:

[0084] The transportation time window matrix is ​​used to describe the serviceable time range of various external logistics resources (such as transportation fleets, loading and unloading points, and supplier scheduling platforms) on different orders or tasks. It is defined as a two-dimensional matrix, including: row index, corresponding to the order number or loading and unloading node; column index, corresponding to the external logistics unit (such as fleet number, loading and unloading point number); cell value, the serviceable time interval (such as [08:30, 09:00]).

[0085] Example: For order O101, the transportation time window matrix T_O101,T1=[10:00,10:30] indicates that fleet T1 can provide transportation services for O101 during this time period.

[0086] A dynamic capacity allocation table is used to represent the capacity arrangement of various external transportation resources at different time periods, including current available capacity, task capacity limit, historical load trends, etc. It is usually modeled with time as the main axis and defined as: time segment (e.g., every 30 minutes); the maximum number of dispatchable times, number of tasks carried, average load, etc. of each fleet or logistics unit within each time period.

[0087] Example: The capacity allocation table for transport fleet T3 is shown below:

[0088]

[0089] Step S5: Based on order feature data, factory internal logistics capacity map, external logistics resource supply strategy and shadow tasks, establish a joint scheduling optimization model based on graph structure modeling to perform collaborative scheduling optimization of production tasks and internal logistics paths;

[0090] The joint scheduling optimization model includes an order-driven production process flow diagram constructed based on order feature data, and an internal logistics network diagram constructed based on the factory's internal logistics capacity diagram. The process flow diagram uses order work units and production equipment as nodes and process dependencies as edges. The internal logistics network diagram uses logistics nodes as nodes and topological connections as edges. The production equipment nodes in the production process flow diagram and the nearest logistics nodes in the internal logistics network diagram are automatically mapped to form a joint graph model. The objective function is set to minimize production task delays and maximize equipment and transportation resource utilization, while the resource transfer rules of the factory's internal logistics capacity diagram and external logistics resource supply strategies are used as boundary conditions for collaborative scheduling optimization.

[0091] When a customized order has not been fully confirmed, but there is a typical configuration cluster generated by configuration intent clustering, shadow tasks are introduced into the joint graph model as pre-scheduled tasks for resource pre-allocation and path pre-scheduling.

[0092] Among them, the order-driven production process flow diagram is a multi-stage task dependency graph built based on the process requirement field in the order feature data. Nodes represent specific processes or required equipment, and edges represent the sequential relationship of processes. For example, order O3 requires "rough milling → finish milling → heat treatment → inspection", which corresponds to the construction of 4 process nodes and their dependent edges.

[0093] The internal logistics network diagram is derived from the internal logistics capacity diagram of the factory constructed in step S3. The logistics nodes in the diagram are abstracted as graph nodes, and the reachable paths between nodes are mapped as edges. At the same time, edge weight attributes (path processing time, channel width level, etc.) are added as a reference for scheduling optimization.

[0094] External logistics resource supply strategies are introduced as input boundary conditions, including a transportation time window matrix and a dynamic capacity allocation table. Their function is to restrict the earliest / latest arrival times of certain tasks and the transportation resources they are allowed to use. For example, order O5 needs to be shipped before 17:00 and can only use fleet T3 allocated by logistics provider L1. This condition will bind the time window constraints and capacity matching between the task node and the logistics exit node in the joint model.

[0095] The joint graph model is a multi-layered heterogeneous graph structure generated by automatically mapping equipment nodes in the process flow diagram to the nearest logistics nodes in the internal logistics diagram. This mapping is generated through nearest neighbor logic (such as minimizing spatial distance and shortest access path) combined with preset rules (such as binding equipment to fixed docking points), ensuring that production tasks can be completely mapped to feasible paths;

[0096] Objective functions include, but are not limited to, combinations of multiple objectives such as minimizing total task delay, maximizing production equipment utilization, minimizing transportation resource conflict rate, and minimizing the number of job batch changes, and support weighted settings or phased optimization;

[0097] In addition to traditional production scheduling constraints (such as non-overlapping equipment and process sequence restrictions), the following key boundary conditions are also introduced: task start window restrictions based on external supply strategies; path capacity and task scheduling cycle time based on logistics capacity maps; and task migration conditions based on resource transfer rules.

[0098] Example: For an order O6 with a process of "assembly → inspection → packaging", if its delivery time is T+6h, and based on the internal logistics capacity diagram, it is known that the "assembly area → inspection area" channel is often at risk of congestion between T+3h and T+4h, then the scheduling optimization model will move the inspection node forward or backward within the allowable range, or adjust the transportation route to avoid high-load sections.

[0099] Step S6: Generate work instructions and transportation scheduling tasks for each production unit and logistics node based on the collaborative scheduling optimization results;

[0100] Among them, the work instruction refers to the specific sequence of actions generated for the production unit, including but not limited to: task number, start and end time of the work, required tooling and process parameters, task priority and delivery deadline, and logistics interface information; the instruction is issued to the equipment control layer or operation terminal through the MES system to drive the production equipment to execute according to the schedule plan.

[0101] For example: "Task T1001, started execution at 08:00 on June 20, equipment number CNC-05, using fixture A23, program path 'P12345.NC', after completion, the material is transferred to loading and unloading port L4";

[0102] Transportation scheduling tasks refer to the instruction plans generated for the dispatch scheduling of in-plant mobile logistics resources (such as AGVs, carts, and forklifts) and external interfaces. The content includes: task number, pickup point and delivery point, allowed window time period, allocated resource number, urgency level label, route preference, etc.

[0103] For example: "AGV scheduling task D233, retrieves items from buffer S5 to loading / unloading point L4, task priority is medium, execution window is 10:00~10:20, recommended path is line-A";

[0104] Work instructions and transportation scheduling tasks are uniformly scheduled and task issued through the scheduling platform, forming a task execution queue. When the external logistics resource supply strategy includes supply window restrictions or scheduling timing suggestions, the system will synchronously embed such constraints into the transportation tasks to ensure overall rhythm matching.

[0105] Supports the joint issuance of task packaging and task splitting strategies: that is, certain job tasks and their corresponding transportation tasks will be packaged and issued as a group of concurrent tasks to avoid equipment idling or path congestion; at the same time, it also allows some large tasks to be split into multiple transportation batches, with scheduling tasks generated separately.

[0106] The scheduling granularity and execution accuracy in this process are limited by the actual availability of current internal logistics resources and system feedback delays. Therefore, the execution system supports dynamic adjustment and rescheduling interfaces to ensure that the scheduling results are adaptable and robust in dynamic environments.

[0107] Step S7: Collect internal logistics information in real time during the execution of the transportation scheduling task to obtain the congestion index of the transportation scheduling task. When the congestion index is abnormal, dynamically adjust the transportation scheduling task.

[0108] The methods for determining abnormal congestion indicators include task congestion and local logistics capacity bottlenecks, and the congestion indicators are calculated in batches at fixed time intervals; the dynamic adjustment includes work order priority, path selection and batch division, and the dynamic adjustment must meet the constraints of the process dependency relationship in the production process flow chart.

[0109] The method for determining task congestion is to obtain the occupancy status of each logistics node in the internal logistics capacity diagram of the factory, calculate the variance of the resource occupancy rate in the occupancy status, and determine that task congestion occurs when the variance of the resource occupancy rate exceeds a set threshold and the duration exceeds a set duration. The method for determining local logistics capacity bottlenecks is to identify logistics nodes in the internal logistics capacity diagram of the factory where resource request blocking occurs continuously, and detect whether the node has a situation where the task waiting time exceeds a set threshold in multiple consecutive scheduling cycles. If so, it is determined to be a local logistics capacity bottleneck.

[0110] The congestion index is calculated based on the internal logistics capacity map of the factory constructed in step S3. The status of each logistics node is abstracted into a Petri net identifier. The following key indicators are calculated through periodic sampling: resource utilization rate, the effective load ratio of the node per unit time; task waiting time, the average waiting time of tasks in the queue; utilization rate fluctuation, the variance of resource utilization rate, used to identify traffic imbalance; and blocking frequency, whether resource request failures occur in multiple consecutive periods.

[0111] Congestion indicators are automatically calculated at fixed scheduling intervals (e.g., 5 minutes). If any indicator exceeds a set threshold, the adjustment module is triggered. Specific adjustments include: adjusting work order priorities, increasing the priority of resource queues for urgent orders, and prioritizing their scheduling; optimizing route selection, switching to logistics channels with lighter loads or backup routes; refactoring batch division, splitting large tasks into multiple smaller batches to reduce instantaneous channel pressure, or merging multiple micro-tasks to reduce handling frequency; and adjusting constraints, all of which must satisfy the process dependencies in the production process flow chart, ensuring that subsequent processes can only be scheduled after the preceding process is completed, to avoid logical disorder.

[0112] Example: The system detects that task T203 has been waiting for more than 12 minutes for 3 consecutive cycles on the assembly → inspection path, and the resource utilization variance has reached the upper limit of the threshold. It is judged as task congestion. The platform's response is to move the inspection process back by 5 minutes and adjust it to the backup path B. At the same time, the priority of task T203 is increased by one level. It is expected that the congestion can be eliminated within 2 cycles.

[0113] Step S8: When the congestion index of the transportation scheduling task is still abnormal after dynamic adjustment, perform path conflict simulation prediction of the transportation scheduling task, and feed the simulation prediction results back to the joint scheduling optimization model to trigger adaptive rescheduling and form a decision closed loop.

[0114] Among them, path conflict simulation prediction refers to constructing a time-driven simulation model based on the current transportation scheduling plan, the status of internal logistics resources and path topology, and predicting and analyzing potential problems such as path conflicts, equipment resource contention and task backlog that may occur in multiple future scheduling cycles.

[0115] Simulation inputs include: the current set of tasks to be executed and their time windows (from step S6); the internal logistics capacity diagram of the factory (Petri net structure); the current status and historical passage statistics of internal logistics nodes (from step S7); and the shipping rhythm and window limits issued by the external supply strategy.

[0116] The predictive model employs event-driven scheduling simulation (Discrete-Event Simulation), with a simulation granularity accurate to the 5-second level of transportation scheduling, for early identification:

[0117] Path conflict: such as two vehicles requesting the same route segment at the same time;

[0118] Channel congestion: Accumulated tasks lead to continuous congestion of logistics channels exceeding the tolerance period;

[0119] Blocking loop: Due to unreasonable path allocation, tasks cross and exclude each other, forming a locked state;

[0120] Feedback mechanisms include:

[0121] If the conflict is unavoidable, a "high-risk path set" and a "task rescheduling suggestion list" will be automatically generated.

[0122] The system inputs this feedback information into the joint scheduling optimization model, which triggers "adaptive rescheduling" as a constraint enhancement or penalty term;

[0123] The joint model automatically regenerates job plans and scheduling paths based on the current scheduling status, boundary conditions, and feedback information;

[0124] Rescheduling optimization maintains the original scheduling dependencies, focusing on adjusting variables such as conflict paths, resource utilization density, and task start and end times to ensure that the goal is achieved with minimal changes based on the original plan.

[0125] Example: The simulation system detected that tasks T402 and T403 are both scheduled to pass through logistics node L7 between 9:00 and 9:10, and only one forklift is available. If they are executed along the original path, traffic congestion will occur. The system generates a path conflict warning and suggests changing the path of T403 to "L5→L8→L9", adjusting the start time to be delayed by 3 minutes, and then issuing a new scheduling task after inputting the joint scheduling model for optimization.

[0126] Example 2, a production optimization system integrating logistics management, see [link to example]. Figure 1 As shown, it includes the following modules:

[0127] Personalized order parsing module: used to receive personalized customer orders and extract order feature data;

[0128] Custom Intent Clustering Module: Used to obtain customer configuration intents before customized orders are fully confirmed, cluster them, generate typical configuration clusters, build corresponding shadow tasks, and manage the introduction, activation, and release of shadow tasks;

[0129] Logistics Status Acquisition Module: Used to collect status information of internal and external logistics resources in the factory, and generate internal logistics information and external logistics information;

[0130] Internal logistics modeling module: used to construct a factory internal logistics capacity map based on internal logistics information, and to describe the topology and resource transfer rules of logistics nodes using Petri net modeling tools;

[0131] External logistics optimization module: This module models factories, logistics suppliers, and their subordinate transportation fleets as intelligent agents, constructs a dynamic game model, and generates external logistics resource supply strategies.

[0132] Supply strategy management module: used to express the supply strategy of external logistics resources in a structured way, and realize its timed update through the communication interface, and generate boundary conditions for the joint scheduling model to call;

[0133] The joint scheduling optimization module establishes a joint scheduling optimization model based on order characteristic data, internal factory logistics capacity map, external logistics resource supply strategy and shadow tasks, and performs collaborative scheduling optimization of production tasks and internal logistics paths.

[0134] The scheduling and execution module generates work instructions and transportation scheduling tasks for each production unit and logistics node based on the results of collaborative scheduling optimization.

[0135] The congestion monitoring and simulation feedback module is used to monitor whether the congestion indicators of transportation scheduling tasks are abnormal, and when the abnormality persists, it performs path conflict simulation prediction and feeds the simulation results back to the joint scheduling optimization module to trigger rescheduling operations.

[0136] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A production optimization method combining logistics management, characterized in that, include: Receive personalized customer orders and extract order feature data; Personalized customer orders for customized products are called customized orders. Before the actual customized order is fully confirmed, the customer's configuration intention is obtained and cluster analysis is performed to generate typical configuration clusters, which are used to construct corresponding shadow tasks. The shadow tasks are introduced into the joint scheduling optimization model as pre-scheduling inputs. Real-time collection of status information of internal and external logistics resources in the factory, referred to as internal logistics information and external logistics information respectively, and modeling of internal logistics information to generate internal logistics capacity map of the factory; The generated internal logistics capacity diagram of the factory includes: A Petri net modeling tool is used to construct an internal logistics capacity map of the factory, which includes the topological connections between logistics nodes and resource transfer rules. The internal logistics information includes the occupancy status and service capabilities of logistics nodes. Logistics nodes include fixed facilities, mobile devices and external interfaces within the factory. The resource transfer rules are constructed based on the service capabilities of each logistics node, including task processing cycle time, path accessibility status and reachability conditions. The factory, logistics supplier and its subordinate external transportation fleet are modeled as intelligent agents. Through a dynamic game model built on a multi-agent reinforcement learning framework, a hierarchical game strategy is adopted to generate the optimal response strategy for each intelligent agent and output the external logistics resource supply strategy, including transportation route allocation, vehicle scheduling, supply time window and task acceptance relationship. Based on order characteristic data, internal factory logistics capacity map, external logistics resource supply strategy and shadow tasks, a joint scheduling optimization model based on graph structure modeling is established to perform collaborative scheduling optimization of production tasks and internal logistics paths. Based on the results of collaborative scheduling optimization, operation instructions and transportation scheduling tasks are generated for each production unit and logistics node. During the execution of transportation scheduling tasks, internal logistics information is collected in real time to obtain congestion indicators for transportation scheduling tasks. When the congestion indicators are abnormal, the transportation scheduling tasks are dynamically adjusted. When the congestion index of the transportation scheduling task is still abnormal after dynamic adjustment, the path conflict of the transportation scheduling task is simulated and predicted, and the simulation prediction results are fed back to the joint scheduling optimization model to trigger adaptive rescheduling and form a decision closed loop.

2. The production optimization method combining logistics management according to claim 1, characterized in that, The strategy for generating the optimal response for each agent using a hierarchical game strategy includes: The factory, logistics supplier and its subordinate external transportation fleet are modeled as intelligent agents for reinforcement learning. Each intelligent agent constructs an observation state set based on order feature data and external logistics information, and pre-sets a set of corresponding optional actions based on its role definition. A dynamic game model is constructed, in which the factory acts as the leader agent to generate the leader's strategy; the logistics supplier and its subordinate external transportation fleet act as follower agents, and after acquiring the leader's strategy, they use reinforcement learning to optimize their own response strategy. The optimization process is based on a set reward function, which includes a weighted combination of transportation cost, on-time delivery rate, and route load balancing. The optimization result is output as an external logistics resource supply strategy.

3. The production optimization method combining logistics management according to claim 2, characterized in that, The external logistics resource supply strategy is expressed in a structured form as a transportation time window matrix and a dynamic capacity allocation table, which serve as the input boundary conditions for the joint scheduling optimization model and are updated in real time through the agent communication interface.

4. The production optimization method combining logistics management according to claim 1, characterized in that, The joint scheduling optimization model includes an order-driven production process flow diagram constructed based on order feature data and an internal logistics network diagram constructed based on the factory's internal logistics capacity diagram; the production process flow diagram uses order work units and production equipment as nodes and process dependencies as edges. The internal logistics network diagram uses logistics nodes as nodes and topological connections as edges; The production equipment nodes in the production process flow diagram are automatically mapped to the nearest logistics nodes in the internal logistics network diagram to form a joint graph model; the objective function is set to minimize production task delays and maximize equipment and transportation resource utilization, while the resource transfer rules of the internal logistics capacity diagram and the external logistics resource supply strategy are used as boundary conditions for collaborative scheduling optimization. When a customized order has not been fully confirmed, but there is a typical configuration cluster generated by configuration intent clustering, shadow tasks are introduced into the joint graph model as pre-scheduled tasks for resource pre-allocation and path pre-scheduling.

5. The production optimization method combining logistics management according to claim 4, characterized in that, The generation and scheduling of the shadow tasks include: Before a customized order is fully confirmed, the customer's configuration intent is constructed as a vector and clustered to generate representative typical configuration clusters. For each configuration cluster, a typical process flow and resource requirement template are generated, and shadow tasks are constructed accordingly. The shadow tasks are introduced into the joint scheduling optimization model in the form of soft constraints to achieve resource pre-allocation and path pre-scheduling before the customized order is confirmed. When the actual order is placed, the corresponding shadow task is matched and activated. Unmatched shadow tasks are released after a set time window to improve the adaptability to the diversity and uncertainty of customized orders. The configuration intent includes customer configuration behavior data on the customization platform and historical order data.

6. The production optimization method combining logistics management according to claim 1, characterized in that, The methods for determining abnormal congestion indicators include task congestion and local logistics capacity bottlenecks, and the congestion indicators are calculated in batches at fixed time intervals; the dynamic adjustment includes work order priority, path selection and batch division, and the dynamic adjustment must meet the constraints of the process dependency relationship in the production process flow chart.

7. The production optimization method combining logistics management according to claim 6, characterized in that, The method for determining task congestion is to obtain the occupancy status of each logistics node in the internal logistics capacity diagram of the factory, calculate the variance of the resource occupancy rate in the occupancy status, and determine that task congestion occurs when the variance of the resource occupancy rate exceeds a set threshold and the duration exceeds a set duration. The method for determining local logistics capacity bottlenecks is to identify logistics nodes in the internal logistics capacity diagram of the factory where resource request blocking occurs continuously, and detect whether the node has a situation where the task waiting time exceeds a set threshold in multiple consecutive scheduling cycles. If so, it is determined to be a local logistics capacity bottleneck.

8. A production optimization system integrating logistics management, characterized in that, The system applies a production optimization method combining logistics management as described in any one of claims 1 to 7, comprising: Personalized order parsing module: used to receive personalized customer orders and extract order feature data; Custom Intent Clustering Module: Used to obtain customer configuration intents before customized orders are fully confirmed, cluster them, generate typical configuration clusters, build corresponding shadow tasks, and manage the introduction, activation, and release of shadow tasks; Logistics Status Acquisition Module: Used to collect status information of internal and external logistics resources in the factory, and generate internal logistics information and external logistics information; Internal logistics modeling module: used to construct a factory internal logistics capacity map based on internal logistics information, and to describe the topology and resource transfer rules of logistics nodes using Petri net modeling tools; External logistics optimization module: This module models factories, logistics suppliers, and their subordinate transportation fleets as intelligent agents, constructs a dynamic game model, and generates external logistics resource supply strategies. Supply strategy management module: used to express the supply strategy of external logistics resources in a structured way, and realize its timed update through the communication interface, and generate boundary conditions for the joint scheduling model to call; The joint scheduling optimization module establishes a joint scheduling optimization model based on order characteristic data, internal factory logistics capacity map, external logistics resource supply strategy and shadow tasks, and performs collaborative scheduling optimization of production tasks and internal logistics paths. The scheduling and execution module generates work instructions and transportation scheduling tasks for each production unit and logistics node based on the results of collaborative scheduling optimization. The congestion monitoring and simulation feedback module is used to monitor whether the congestion indicators of transportation scheduling tasks are abnormal, and when the abnormality persists, it performs path conflict simulation prediction and feeds the simulation results back to the joint scheduling optimization module to trigger rescheduling operations.

Images