An intelligent order scheduling method and system for a preliminary order of a clothing industry based on operations research

Abstract

The application discloses a kind of based on operations' clothing industry initial edition order intelligent scheduling method and system, belong to supply chain intelligent production scheduling technical field.Method includes: acquisition demand, capacity and supplier pool classification information, and the supplier is divided into core pool and qualified pool, and the capacity is subdivided into six dimensions of day-level capacity unit of supplier, brand, channel, age, face kind, variety subdivided pool, and each unit is defined as algorithm supplier;Based on dimension matching relationship definition private and common capacity set and common priority;Build mixed integer programming model containing hard constraint (capacity, label matching, single bundle, delivery date etc.) and optimization target (minimum delay, capacity achievement, even depth balance, common priority etc.);Call solver and output each demand allocation algorithm supplier and daily plan.The application realizes multi-dimensional day-level granularity global optimal scheduling, introduces flexible capacity sharing and even depth balance mechanism, improves supply chain efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent production scheduling technology in the apparel industry supply chain, specifically to an intelligent scheduling method and system for initial orders in the apparel industry based on operations research. Background Technology

[0002] The order scheduling process in the apparel industry is divided into initial order scheduling, color-matching review, and bulk order scheduling. Initial order scheduling is a crucial decision-making process that rationally allocates orders to suppliers based on virtual demand derived from product planning and market forecasts. Existing order scheduling systems primarily rely on rule engines or human experience, exhibiting the following shortcomings: order scheduling rules are fragmented and rigid, lacking a unified mathematical modeling framework; multi-dimensional constraints (capacity, brand, channel, process labels, etc.) cannot be collaboratively optimized; capacity management is limited to a supplier-monthly dimension, lacking sufficient granularity; the lack of a deep balancing mechanism leads to order fragmentation; and order scheduling results lack mathematical theoretical support, making it difficult to guarantee global optimization. Therefore, an intelligent order scheduling method based on operations research optimization is urgently needed. Summary of the Invention

[0003] The technical problem to be solved by this invention is to overcome the shortcomings of existing technologies, such as scattered order scheduling rules, inability to unify modeling, difficulty in coordinating and optimizing multiple constraints, extensive capacity management, lack of average depth balancing, and reliance on human experience, and to provide a method and system for intelligent order scheduling of initial orders in the apparel industry based on operations research.

[0004] The technical solution adopted in this invention is as follows:

[0005] A method for intelligent order scheduling in the apparel industry based on operations research, comprising:

[0006] Step 1: Data Acquisition and Preprocessing: Obtain initial demand data, supplier capacity data, and supplier pool classification information. Divide suppliers into a core supplier pool and a qualified supplier pool. Further subdivide supplier capacity into daily-level capacity units, encompassing at least six dimensions: supplier, brand, channel, age group, noodle type, and product variety. Each daily-level capacity unit is defined as an algorithmic supplier. This step ensures the multi-dimensional, detailed capacity data and supplier classification foundation required for subsequent modeling, providing data support for global optimization.

[0007] Step 2: Definition of Capacity Sharing Mechanism: Based on the matching relationship between demand and algorithm suppliers across the six dimensions, define the private capacity set and the shared capacity set, and set priority targets for capacity sharing. This mechanism solves the problem of flexible allocation when private capacity is insufficient, and improves capacity utilization.

[0008] Step 3: Model Construction: A mathematical model is constructed based on mixed-integer programming. This model includes hard constraints and an optimization objective function. Hard constraints include at least supplier capacity constraints, demand-algorithm supplier label matching constraints, single-item bundling constraints, and order delivery date constraints. The optimization objective function includes at least the objectives of minimizing delayed orders, achieving capacity fulfillment rate, achieving deep balancing of supplier orders, and prioritizing shared capacity. Through unified modeling, collaborative optimization with multiple constraints and objectives is achieved.

[0009] Step 4: Solving and Outputting: Use a mixed-integer programming solver (such as COPT) to solve the mathematical model, and convert the solution into a scheduling result that includes the algorithmic supplier for each demand allocation and the daily production plan. This step ensures the global optimality and executability of the scheduling result.

[0010] Furthermore, the hard constraints may also include at least one of the following: supplier quantity constraints, continuous production constraints, manual order allocation constraints, PK product concentration constraints, non-special order quantity constraints, best-selling product constraints, cancellation of demand allocation constraints, suspension of supplier allocation constraints, and online / offline time constraints.

[0011] Furthermore, the objective of balancing the average order depth of suppliers is incorporated into the mathematical model through linear approximation. Deviation variables are introduced for both the core supplier pool and the qualified supplier pool to penalize the deviation between the order depth within each pool and the overall order depth. This process transforms the nonlinear problem into a linear one, ensuring the model's solvability and effectively mitigating order fragmentation.

[0012] Furthermore, the specific objectives for prioritizing shared production capacity are as follows: first, to use private production capacity that fully matches demand across the six dimensions; second, to use production capacity that matches the type of noodles but spans different product sub-pools or age groups; third, to use production capacity across channels; and finally, to use production capacity across brands.

[0013] Furthermore, the solution and output steps can adopt a hierarchical sequence method: the first solution stage aims to minimize the number of delayed orders and determine the maximum set of achievable demands; the second solution stage fixes this set of demands and then optimizes objectives such as capacity achievement rate and average depth balance. This strategy achieves phased collaboration of multiple objectives.

[0014] This invention also provides an intelligent order scheduling system for the initial order of the apparel industry based on operations research, which implements the aforementioned intelligent order scheduling method for the initial order of the apparel industry. The system includes a data preprocessing module, a model building module, a solution module, and a result generation module, which respectively implement the functions of each step of the above method.

[0015] Compared with the prior art, the advantages of this invention are as follows:

[0016] (1) The scheduling rules are uniformly modeled as MIP to achieve global optimal solution, which overcomes the limitation of traditional rule engines that only obtain local feasible solutions;

[0017] (2) Modeling at a six-dimensional, day-level granularity and introducing a flexible capacity sharing mechanism to solve the problems of extensive capacity management and low utilization rate;

[0018] (3) By incorporating the average depth balance into the model through linear approximation, the uneven supplier load caused by order fragmentation is effectively solved;

[0019] (4) The hierarchical sequence method is adopted to make the scheduling results quantifiable, traceable and configurable. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein:

[0021] Figure 1 This is a flowchart illustrating the intelligent order scheduling method for the initial version of the apparel industry based on operations research in Embodiment 1 of the present invention.

[0022] Figure 2 This is a block diagram of the module composition of the initial version of the intelligent order scheduling system for the apparel industry based on operations research in Embodiment 2 of the present invention. Detailed Implementation

[0023] The present invention will be further described in detail below with reference to specific embodiments. These embodiments are merely one implementation of the present invention and are not intended to limit the scope of protection. Specific conditions not specified in the embodiments should be performed according to conventional methods or manufacturer recommendations.

[0024] Example 1

[0025] like Figure 1 As shown, this invention provides a method for intelligent order scheduling of initial orders in the apparel industry based on operations research, including:

[0026] Step S1: Data Acquisition and Preprocessing;

[0027] The system first collects the following data from a supply chain management system (such as ERP):

[0028] The initial demand set D contains the following for each demand: style, estimated production capacity requirement (line days) Q_d, earliest start date, latest end date, cooperation mode (ODM / FOB / integrated / CFOB), quantity rating (S / A / B, etc.), and special labels (bestseller, PK style, IP licensing, process requirements, etc.).

[0029] Supplier capacity data: A set of physical suppliers P, each physical supplier contains algorithmic suppliers s (i.e., six-dimensional granularity: supplier-brand-channel-age group-noodle type-product segment pool), and the daily capacity limit C_st and monthly capacity limit Cap_s_month for each algorithmic supplier s at time t.

[0030] Supplier pool classification information: Entity suppliers are divided into a core supplier pool P_core and a qualified supplier pool P_qualified.

[0031] Preprocessing also includes:

[0032] Prioritize needs based on cooperation models (ODM > FOB > Integration > CFOB).

[0033] The earliest launch date and latest production date for each demand are calculated based on the fabric set-up time or the standard fabric production cycle. When using the standard fabric order date for calculation, the order must be produced according to the optimal production line, and the earliest launch date can be advanced by up to 90 days only if the estimated quantity of that item is in the top 70% of the estimated quantities of all items in descending order across the brand-channel-fabric type dimensions.

[0034] In this way, by breaking down and prioritizing daily production capacity across six dimensions, a refined and orderly data foundation is provided for subsequent global optimization, avoiding the inefficiency caused by extensive production capacity management.

[0035] Step S2: Definition of the capacity sharing mechanism;

[0036] Based on the matching relationship between demand d and algorithm supplier s across six dimensions, the following definition is made:

[0037] Private capacity set S_d_match: A set of algorithm suppliers that are perfectly matched with demand d in terms of brand, channel, age group, type of noodle, and variety segmentation pool.

[0038] Shared capacity set S_d_share: A set of non-matching type algorithmic suppliers that can be enabled when private capacity is insufficient, and crossing is only allowed in the age group and variety segment pool dimensions (e.g., cross variety segment or cross age group under the same seed, cross channel if necessary, and finally cross brand).

[0039] At the same time, a priority target Z_share for shared production capacity is set (see the objective function below), and the sharing ratio is controlled by a penalty coefficient to balance the independence of brand channels and the flexible allocation of production capacity.

[0040] In this way, when private capacity is insufficient, the shared capacity pool is automatically activated, and the priority of sharing is quantitatively controlled through the objective function penalty term, which improves the overall capacity utilization rate without violating the core constraints.

[0041] Step S3: Model building;

[0042] A mathematical model is constructed based on mixed integer programming (MIP), and the model includes the following core elements:

[0043] (a) Definition of a set;

[0044] D: Demand set;

[0045] S: Set of algorithm providers (six-dimensional granularity), with each algorithm provider s corresponding to a unique six-dimensional combination;

[0046] P: Collection of physical suppliers;

[0047] M: Single item collection;

[0048] T: Time set (days);

[0049] F: Collection of fabric types (knitted, woven, denim, wool);

[0050] V: A comprehensive collection of all brand, channel, age group, noodle type, and variety segmentation dimensions;

[0051] S_d_match, S_d_share: The private and shared supplier sets for demand d;

[0052] P_core, P_qualified: Core pool and qualified pool.

[0053] (II) Parameter Definition

[0054] Q_d: Capacity demand for demand d (line days);

[0055] C_st: The maximum production capacity of algorithm supplier s on day t;

[0056] Cap_s_month: The monthly capacity limit of algorithm supplier s;

[0057] target_sp: The planned capacity of algorithm supplier s to achieve its target within the specified time period;

[0058] π_pv_target: The target proportion of entity supplier p allocated to segment dimension v;

[0059] N_pk: The maximum number of PK (player kill) contracts allowed without penalty;

[0060] w1~w9: Weight penalty coefficients for each optimization objective.

[0061] (III) Decision Variables

[0062] x_dt: Production quantity of demand d at time t (line days), a continuous variable;

[0063] y_dts: The quantity of demand d produced by algorithm supplier s at time t, a continuous variable;

[0064] z_ds: Demand d is the total amount produced by algorithm supplier s, a continuous variable;

[0065] β_d: Whether demand d will be produced, a 0-1 variable;

[0066] α_mp: Whether a single item m is produced by the physical supplier p, a 0-1 variable;

[0067] δ_core_v, δ_qual_v: The order depth deviation between the core pool and the qualified pool in dimension v, continuous variables;

[0068] ε_p_pk: The deviation between the quantity of PK items accepted by supplier p and the maximum allowable quantity, a continuous variable.

[0069] (iv) Hard constraints

[0070] This invention establishes the following hard constraints in the MIP model (some of which can be enabled based on business configuration):

[0071] Supplier quantity constraint: Each demand can be assigned to at most one entity supplier, and partial satisfaction is not supported. Mathematical expression: ∑_s z_ds ≤ Q_d·β_d, β_d∈{0,1}.

[0072] Single-item bundling constraint: All requirements for the same single item m must either be fully fulfilled or none must be produced, and the same supplier group must be used. Mathematical expression: For any requirements d1 and d2 of the same single item m, β_d1 = β_d2, and α_mp is consistent for all requirements.

[0073] Continuous production constraint (configurable): Production continues until completion once demand begins.

[0074] Supplier capacity constraints:

[0075] Daily production capacity: ∑_d y_dts ≤ C_st, ∀s,t;

[0076] Monthly production capacity: ∑_d z_ds_month ≤ Cap_s_month, ∀s, month;

[0077] Manual order allocation constraint: If the demand has a specified physical supplier (such as an ODM factory selection), it can only be allocated to that supplier.

[0078] PK (Primary Target) Constraint: PK pairs (m1, m2) must be allocated to the same entity supplier: α_m1p = α_m2p, ∀p.

[0079] Non-special order quantity constraints: Quantity requirements with an S or A rating and no special labels can only be allocated to core pool suppliers.

[0080] Bestseller Constraint: Products with the bestseller tag can only be allocated to core pool suppliers unless they are reordered.

[0081] Cancel requirement style assignment constraint: Canceled requirements are removed from D.

[0082] Suspend supplier allocation constraints: Suspended suppliers are removed from S.

[0083] Production start and end date constraints: The production date in the demand simulation must be after the earliest start date and before the latest end date.

[0084] Supplier constraints: Physical suppliers must match all relevant tags of the requirements (IP licensing, process, brand licensing, channels, dough type capability, variety segmentation pool).

[0085] (v) Optimize the objective function

[0086] Multi-objective function in weighted sum form:

[0087] Min Z = w1·Z1 + w2·Z2 + w3·Z3 + w4·Z4 + w5·Z5 + w6·Z6 + w7·Z7+ w8·Z8 + w9·Z9;

[0088] The sub-objectives are defined as follows:

[0089] Z1, Pre-arranged Supplier Priority: Z1 = ∑_{m,p≠p_pre} α_mp·Q_m, encourages prioritizing suppliers in the sample garment stage to reduce order transfer costs.

[0090] Z2, Capacity Planning Achievement Rate: Z2 = ∑_{s,month} |actual_s_month - target_s_month|, ensuring that actual capacity closely matches the planned target. The target value is adjusted based on the order arrival progress gradient to prevent order scheduling instability when there are few orders.

[0091] Z3, the minimum number of delayed orders: Z3 = w3a·∑(1-β_d) + w3b·∑(1-β_d)·Q_d, which penalizes both the number of delayed orders and the amount of delayed goods.

[0092] Z4, Low-pass fabric supplier centralization: Distribute the demand for the same type of low-pass fabric to the same entity supplier as much as possible to reduce procurement and logistics costs (configurable and can be enabled).

[0093] Z5, Supplier Order Depth Balancing: For each brand-channel-fabric-age-variety segment pool dimension, the deviation between the average depth of the core pool and the average depth of the qualified pool and the overall average depth is calculated separately. Deviation variables δ_core_v and δ_qual_v are introduced for linear penalty. The overall average depth depth_v = (∑_{m∈M_v} estimated) / |M_v|. The deviation between the core pool average depth and depth_v is approximated by linear constraints; the larger the deviation, the greater the penalty.

[0094] Z6, PK (Primary Transaction) Distribution Objective: Minimize the number of PK projects undertaken by a single supplier. Set a maximum allowed number of PK projects N_pk without penalty, and penalize any exceeding this limit ε_p_pk.

[0095] Z7, Supplier Priority Category Target: Encourage suppliers to take on categories they are good at, thereby improving production efficiency and quality.

[0096] Z8, Supplier Category Fitting Target: To ensure that the actual allocation ratio of suppliers in the brand-product segmentation pool aligns with the target ratio determined during the capacity planning phase. Z8 = max_p(∑_v |π_pv - π_pv_target|).

[0097] Z9, SA style even distribution target: to match the number of S and A styles undertaken by suppliers in the brand-type dimension with the production capacity ratio, and to avoid excessive concentration or dispersion of high-quality styles.

[0098] In this way, the 9 objectives cover multiple dimensions of KPIs, from delivery time, capacity, cost to quality balance. By adjusting the weights, they can be flexibly adapted to different business priorities to achieve optimal global synergy.

[0099] Step S4: Solve and output.

[0100] Commercial solvers (such as COPT) are used to solve the above MIP model. To improve solution efficiency, the following strategy is adopted:

[0101] Variable preprocessing: Before model building, for (d,s,t) combinations with mismatched labels or non-overlapping time windows, the corresponding decision variable y_dts is not created, which greatly reduces the number of variables.

[0102] Hierarchical sequence method (as a preferred implementation):

[0103] Phase 1: Solve the problem with the sole objective of minimizing ∑(1-β_d) (i.e. the number of delayed orders) to determine the maximum set of demand that can be fulfilled.

[0104] The second stage involves fixing the optimal β_d result from the first stage (allowing for slight relaxation to cope with floating-point errors) and then optimizing the remaining objectives (capacity achievement rate, average depth balance, etc.).

[0105] The penalty coefficient is dynamically adjusted: the weights of each target, w1 to w9, can be flexibly configured according to business priorities, and different weight combinations are used in different scheduling scenarios.

[0106] After the solution is obtained, the optimal solution is converted into a standardized scheduling result, which includes:

[0107] Each demand allocation includes the entity supplier and the specific algorithm supplier (six dimensions).

[0108] Daily production plan (line days);

[0109] Capacity utilization details (by day, by supplier);

[0110] Various KPI indicators (delay rate, capacity achievement rate, average depth, PK model dispersion, etc.).

[0111] Once the scheduling results are confirmed by the business personnel, they can be locked. Locked requests will maintain their supplier allocation and capacity utilization unchanged in subsequent scheduling iterations.

[0112] In this way, through variable preprocessing and hierarchical sequence method, even when facing large-scale real-world scheduling problems (thousands of demands, hundreds of suppliers), a high-quality global optimal solution can be obtained within an acceptable time; the scheduling results are quantifiable and traceable, supporting business iteration locking.

[0113] Example 2

[0114] This embodiment 2 provides an initial version of an intelligent order scheduling system for the apparel industry based on operations research, including a data preprocessing module 1, a model building module 2, a solution module 3, and a result generation module 4.

[0115] The data preprocessing module 1 is used to obtain initial demand data, supplier capacity data and supplier pool classification information, divide suppliers into core supplier pool and qualified supplier pool, and subdivide supplier capacity into six-dimensional day-level capacity units, each unit being defined as an algorithm supplier.

[0116] The model building module 2 is used to build a mathematical model based on mixed integer programming, including the aforementioned hard constraints and optimization objective function; and is used to define the private capacity set and the shared capacity set based on the matching relationship between demand and algorithm suppliers in the six dimensions, and set the priority target for capacity sharing.

[0117] The solution module 3 is used to call a mixed integer programming solver (such as COPT) to solve the mathematical model, preferably using the hierarchical sequence method.

[0118] The result generation module 4 is used to convert the solution results into a scheduling result that includes the algorithm supplier for each demand allocation and the daily production plan.

[0119] Example 3

[0120] A clothing brand company needs to schedule approximately 500 initial design requirements (styles) at the beginning of each quarter, involving 20 physical suppliers. Each supplier is further subdivided into an average of 50 algorithmic suppliers (a six-dimensional combination), with a time window of 120 days. Using the method of this invention, approximately 1.2 million candidate assignment variables are generated after data preprocessing, which are then reduced to 300,000 after variable preprocessing. The constructed MIP model contains approximately 150,000 variables (including approximately 5,000 0-1 variables) and 100,000 constraints. The COPT solver is called, employing a hierarchical sequence method. The first stage determines the maximum feasible requirement set in approximately 30 seconds (completion rate 98%), and the second stage completes multi-objective optimization in approximately 2 minutes. The output scheduling results show: only 2 delayed orders, a 22% increase in the average depth of the core pool, an 18% increase in the average depth of the qualified pool, and an increase in capacity utilization from 75% to 92%. After confirmation by business personnel, the results are locked for subsequent color matching review and bulk order scheduling. This case verifies the efficiency and superiority of this invention.

[0121] Industrial applicability

[0122] The method and system of this invention can be deployed in the supply chain management system of apparel companies. By calling the standard MIP solver, automated order scheduling can be achieved, which can significantly improve order scheduling efficiency, reduce labor costs, and optimize the overall performance of the supply chain.

[0123] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent order scheduling in the apparel industry based on operations research, characterized in that, Includes the following steps: Data collection and preprocessing: Obtain initial demand data, supplier capacity data and supplier pool classification information, divide suppliers into core supplier pool and qualified supplier pool, and subdivide supplier capacity into daily-level capacity units with at least six dimensions including supplier, brand, channel, age group, noodle type and variety. Each daily-level capacity unit is defined as an algorithm supplier. Capacity sharing mechanism definition: Based on the matching relationship between demand and algorithm suppliers in the six dimensions, define the private capacity set and the shared capacity set, and set the capacity sharing priority target; Model Construction: A mathematical model is constructed based on mixed integer programming. The mathematical model includes hard constraints and an optimization objective function. The hard constraints include at least supplier capacity constraints, demand-algorithm supplier label matching constraints, single-item bundling constraints, and order delivery date constraints. The optimization objective function includes at least the objectives of minimizing delayed orders, achieving capacity completion rate, achieving deep balance of supplier orders, and prioritizing shared capacity. Solution and Output: Call the mixed integer programming solver to solve the mathematical model, and convert the solution into a scheduling result that includes the algorithm supplier for each demand allocation and the daily production plan.

2. The intelligent order scheduling method for initial orders in the apparel industry based on operations research as described in claim 1, characterized in that, The hard constraints also include at least one of the following: supplier quantity constraints, continuous production constraints, manual order allocation constraints, concentrated constraints on PK products, non-special order quantity constraints, best-selling product constraints, cancellation of demand allocation constraints, suspension of supplier allocation constraints, and online / offline time constraints.

3. The intelligent order scheduling method for initial orders in the apparel industry based on operations research as described in claim 1, characterized in that, The objective of balancing the average order depth of suppliers is incorporated into the mathematical model through linear approximation. Specifically, this includes introducing a deviation variable for both the core supplier pool and the qualified supplier pool to penalize the deviation between the order depth in each supplier pool and the overall order depth.

4. The intelligent order scheduling method for initial orders in the apparel industry based on operations research as described in claim 1, characterized in that, The specific priorities for shared production capacity include: prioritizing the use of private production capacity that fully matches demand across the six dimensions; secondly, using production capacity that matches the type of noodle but spans different product sub-pools or age groups; thirdly, using production capacity across different channels; and finally, using production capacity across different brands.

5. The intelligent order scheduling method for initial orders in the apparel industry based on operations research as described in claim 1, characterized in that, The solution and output steps employ a hierarchical sequence method, specifically including: The first solution stage: The goal is to minimize the number of delayed orders and determine the maximum set of requirements that can be fulfilled. Second solution stage: Fix the result of the maximum achievable demand set obtained in the first solution stage, and then solve the optimization objective function that includes the capacity achievement rate target and the deep balance target of supplier orders.

6. The intelligent order scheduling method for initial orders in the apparel industry based on operations research as described in claim 1, characterized in that, The data collection and preprocessing steps also include: determining the priority of demand based on the cooperation mode of the style, and calculating the earliest online date and the latest offline date of each demand based on the fabric set completion time or the standard fabric production cycle.

7. The intelligent order scheduling method for initial orders in the apparel industry based on operations research as described in claim 1, characterized in that, The demand-algorithm supplier tag matching constraint includes matching verification of at least one of the following tags: IP authorization tag, process tag, brand authorization tag, channel tag, noodle variety capability tag, and variety segmentation pool tag.

8. A system for intelligent scheduling of initial orders in the apparel industry based on operations research, used to execute the intelligent scheduling method for initial orders in the apparel industry based on operations research as described in any one of claims 1-7, characterized in that, include: The data preprocessing module is used to obtain initial demand data, supplier capacity data and supplier pool classification information, divide suppliers into core supplier pool and qualified supplier pool, and subdivide supplier capacity into daily-level capacity units with at least six dimensions including supplier, brand, channel, age group, noodle type and variety subdivision pool. Each daily-level capacity unit is defined as an algorithm supplier. The model building module is used to construct a mathematical model based on mixed-integer programming. The mathematical model includes hard constraints and an optimization objective function. The hard constraints include at least supplier capacity constraints, demand-algorithm supplier label matching constraints, single-item bundling constraints, and order delivery date constraints. The optimization objective function includes at least the objectives of minimizing delayed orders, achieving capacity completion rate, achieving deep balance of supplier orders, and prioritizing capacity sharing. Furthermore, the model building module is also used to define a private capacity set and a shared capacity set based on the matching relationship between demand and algorithm suppliers in the six dimensions, and to set the priority objective for capacity sharing. The solver module is used to call the mixed integer programming solver to solve the mathematical model; The results generation module is used to convert the solution results into scheduling results that include the algorithm supplier for each demand allocation and the daily production plan.

9. The intelligent order scheduling system for the initial version of garment industry based on operations research as described in claim 8, characterized in that, The solution module uses a hierarchical sequence method for solving the problem: first, it solves for the maximum set of demand that can be completed with the goal of minimizing the number of delayed orders, and then fixes the maximum set of demand that can be completed; then, based on the fixed set of demand, it optimizes the objective function that includes the target of capacity achievement rate and the target of deep balance of supplier orders.

Images