A complex production design planning method based on RSOPT model

Abstract

This invention discloses a complex production design and planning method based on the RSOPT model, belonging to the field of complex product design and production planning technology. Addressing the problems of traditional RSOP models such as the lack of a time dimension, the inability of static planning to adapt to design iterations and operational delays, and resource waste caused by kitting waiting, this invention introduces a time parameter T on top of the RSOP quadruple, constructing an RSOPT quintuple production design unit. It establishes a structure matrix, a diagonal duration matrix, and an external gate control vector, using maximum-addition algebra to solve for the earliest start and completion times of each unit, quantifying the kitting waiting time. By monitoring parameter changes in real time, it achieves closed-loop adaptive adjustment by partially recalculating affected downstream units, and combines duration optimization, gating adjustment, and structural reconstruction to achieve planning optimization. This method can effectively handle time-dependent and uncertain events in complex production, reduce kitting waiting and schedule risks, and improve the dynamism, robustness, and resource utilization efficiency of production design and planning. It is applicable to complex product production design scenarios such as aerospace and shipbuilding.

Description

Technical Field

[0001] This invention relates to a complex production design and planning method based on the RSOPT model. Background Technology

[0002] In the context of digital transformation, complex production design and planning fields (such as aerospace and shipbuilding) face dynamic challenges in design paths. Existing RSOP models primarily focus on static resource cost calculation, decomposing the production process into resource, site, operation, and product unit structures. The output product of each unit is the input resource of the next unit, with resources equivalent to products. Production is achieved through continuous unit RSOP operations. However, this model ignores the dynamic impact of the time dimension, such as design iteration time windows, operational delays, and demand fluctuations, resulting in insufficient adaptability in complex product design environments. Traditional design planning methods rely on CAD / PLM systems and static simulations, which cannot effectively handle sudden changes, leading to low design efficiency and resource waste. Summary of the Invention

[0003] This invention aims to overcome the static limitations of existing RSOP models and provide a complex production design planning method based on the RSOPT model. This method includes: constructing RSOPT quintuple task units and generating a dynamic continuous cRSOPT network through derivation; applying time series forecasting algorithms to optimize parallel design paths; and real-time monitoring and adjustment of T parameters to form a closed-loop feedback continuity. Compared to existing technologies, the novelty of this invention lies in the introduction of dynamic expansion of T parameters, the inventiveness in the integration of time series algorithms and the RSOP model, and the practicality in its applicability to complex product design planning, reducing the risk of time schedule iterations in design planning.

[0004] The technical solutions adopted in this invention are as follows:

[0005] A complex production design and planning method based on the RSOPT model, wherein the RSOPT model is represented by a five-tuple consisting of resources, site, operation, product and time parameters to represent the production design unit, and the time parameters include start time, end time and kitting waiting time;

[0006] The method includes the following steps:

[0007] The production design process is decomposed into several five-tuple units, and the resources, site, operation, product and process duration of each unit are determined.

[0008] A structural matrix is ​​established based on the inter-unit dependencies, a diagonal duration matrix is ​​established based on the process duration, and an external gate control vector is established based on external time constraints.

[0009] Based on the maximum-addition algebra, the start and end times of each unit are calculated according to the structure matrix, diagonal duration matrix and external gate control vector;

[0010] For units with multiple preceding inputs, the kitting waiting time is calculated based on the start time and the end time of each preceding unit.

[0011] Furthermore, the process duration includes a baseline duration and a disturbance term. The baseline duration is determined based on the process specifications and historical data, and the disturbance term is determined based on time series prediction or real-time monitoring.

[0012] Furthermore, the resources include raw materials, work-in-process, machine time, labor time, fixture positions, energy quotas, and buffer positions, and the resources have category identifiers, quantities, and available time window attributes;

[0013] The site includes a physical site, a virtual site, or a hybrid site, and the site has capacity constraints and parallelism attributes;

[0014] The operation includes steps, processes, or processing methods, and the operation has standard working hours, efficiency indicators, and complexity indicators.

[0015] The products include intermediate or final results, and the products have attributes of type, quality and availability time. The output product of the previous unit serves as the input resource for the next unit.

[0016] Furthermore, determine the resources, site, operation, product, and process duration for each unit, including:

[0017] Obtain historical project structure data, resource inventory logs, process specifications, and Gantt chart time logs from CAD, PLM, ERP, or WMS systems;

[0018] The quintuple unit is generated based on the data.

[0019] Furthermore, in the structure matrix, the corresponding element takes a zero value when there is a dependency relationship between units, and takes a negative infinity value when there is no dependency relationship.

[0020] Furthermore, the computation based on maximum-addition algebra includes:

[0021] The maximum value operation is used as the addition operation, and ordinary addition is used as the multiplication operation;

[0022] The starting time vector is determined based on the result of the multiplication of the structure matrix and the end time vector and the result of the addition of the external gate control vector.

[0023] The end time vector is determined based on the result of the multiplication operation between the diagonal duration matrix and the start time vector.

[0024] The calculation of the kitting wait time for units with multiple preceding inputs includes:

[0025] For each direct preceding unit, calculate the current unit's start time and the preceding unit's end time according to the formula: Calculate the waiting time corresponding to this pre-processor unit, where This represents the waiting time from the k-th preceding unit in the i-th layer to the current unit. Indicates the start time of the current cell. This represents the end time of the k-th pre-unit in the i-th layer;

[0026] The kitting wait time of this unit is obtained by summing the waiting times of all directly preceding units.

[0027] Furthermore, it also includes: based on the dependencies between units, according to the formula Calculate the total kitting wait time, where Indicates the total waiting time for all items to be completed. Indicates the start time of the downstream unit. This indicates the end time of the upstream unit, and the summation range covers all unit pairs with dependencies.

[0028] Furthermore, it also includes: monitoring changes in the process duration or external time constraints; determining the affected set consisting of the affected units and their downstream reachable units; and recalculating the start time, end time, and kitting wait time only within the affected set in topological order.

[0029] Furthermore, it also includes: identifying the critical path based on the kitting wait time;

[0030] By adjusting the matching relationships of resources, site, operation, or product inputs and outputs to compress the process time of units on the critical path, or by adjusting external time constraints to delay the start time of units on the non-critical path, or by adjusting the dependencies between units to restructure the production process, the kitting wait time or end-of-life completion time can be optimized.

[0031] The present invention has the following beneficial effects:

[0032] (1) By introducing time parameters including start time, end time and kitting waiting time, the static quadruple is extended into a quintuple structure, enabling the production design unit to have the ability to describe time sequence, reflect the impact of design iteration time window, operation delay and demand fluctuation on planning, thereby supporting time-dependent design path optimization.

[0033] (2) In response to the resource waiting phenomenon caused by multiple pre-processing merging in complex production, by establishing an explicit calculation relationship of the set waiting time, the implicit buffer time in traditional planning is transformed into a calculable quantitative indicator, so that the time schedule risk is transformed from experience judgment to numerical evaluation, and provides a data basis for identifying critical paths and optimizing resource allocation.

[0034] (3) By establishing a closed-loop feedback mechanism, when the process duration or external constraints change, the timing parameters can be updated by only performing local recalculation within the affected downstream unit set, avoiding recalculation of the entire network, shortening the calculation time required for dynamic adjustment, and supporting real-time correction of the planning scheme.

[0035] (4) Based on the quantified time parameters, the critical path duration can be compressed by adjusting resources, site, operation or product structure, or work-in-process waiting can be balanced by adjusting external time constraints, or the merging level can be changed by reconstructing unit dependencies, thereby obtaining an optional optimization scheme between time schedule, resource utilization and buffer scale.

[0036] (5) By establishing a unified matrix-vector calculation framework through maximum-addition algebra, the unit dependency relationship, process duration and external constraints are structurally represented to ensure that the earliest feasible start and completion time is obtained under the condition of satisfying the pre-constraints, thereby reducing design rework caused by time conflicts. Attached Figure Description

[0037] Figure 1 This is a diagram of the overall architecture of the present invention.

[0038] Figure 2 This is a flowchart of the present invention. Detailed Implementation

[0039] The invention will now be further described with reference to the accompanying drawings.

[0040] like Figure 1 and Figure 2 This embodiment breaks down the complex production design process in fields such as aerospace and shipbuilding into several independent RSOPT model quintuple production design units according to the sequential logic of production operations. The standardized expression for each unit is as follows: This expression is an extension of the traditional RSOP quadruple. During the decomposition and definition process, it extracts historical project structure data, resource inventory logs, process specification documents, standard time libraries, historical OEE data, and Gantt chart time logs from CAD, PLM, ERP, and WMS systems. Based on this data, it determines the resources, site, operation, product information, and process duration for each unit. Among these:

[0041] Let i be the set of input resources for unit i, expressed as: , This is the set of resource categories for this unit, encompassing raw materials, work-in-process, machine hours, labor hours, jig locations, energy quotas, and buffer locations. The attribute format for a single resource component is {id, type, quantity, availability}, and data is sourced from existing ERP / BOM / PLM / WMS system resource inventory logs, etc. Input is estimated from historical design project data or selected based on site conditions (S), iterating until site constraints are matched. The expression for the resource availability time window is: .

[0042] The work site for unit i can be a physical site, a virtual site, or a hybrid site, with the attribute format {id, name, constraints, capacity}. This invention obtains the site settings through a forward derivation or reverse decomposition process, with one site determining one unit. Specifically, the overall design process can be broken down into independent calculation units, initially derived from project specifications, iterating to minimize constraint conflicts. One site determines the capacity, constraints, and parallelism.

[0043] The operation process and technology of unit i include steps, procedures, processing methods and their corresponding standard working hours, equipment overall efficiency (OEE), and complexity index. The attribute format is {id, method, efficiency, complexity}. The data sources are process specifications, MTM / standard working hour library, and historical OEE.

[0044] The output product of unit i after the production operation includes intermediate and final results in the production design process. Its attribute format is {id, type, quality, availability}. The output product of the previous unit... Directly used as input resources for the next unit. The two formats are completely matched, and only the input resources of the initial unit and the output products of the final unit are not associated with other units.

[0045] For unit The time parameter is expressed using a triplet expression. .in The start time of the unit is determined by aligning the completion time of the precursor unit with the changeover and transfer time; The unit completion time is determined by the processing time and the delay term. The waiting time for downstream alignment, also known as the fitting wait time, has a value range of [value missing]. .

[0046] In complex production scenarios, different components may have different process schedules, meaning that different upstream units are unlikely to complete production simultaneously. The outputs of these units must be used as resources to input into the next unit, leading to resource waiting times for subsequent units, also known as "kiting wait." Therefore, not all upstream units... All are strictly equal to the subsequent units. There is a waiting time while all resources are ready. If there is no need to wait, then =0.

[0047] The key point of this invention is to calculate the waiting time, thereby redesigning the units and links to obtain the shortest time for the complete project.

[0048] + = ,

[0049] For unit Baseline duration calculated based on process and capacity:

[0050] ,

[0051] In actual production, if there are predicted intra-cell delay terms... (From time series prediction / feedback monitoring), then the actual duration of the unit. :

[0052] .

[0053] unit → Indicates from the first Laminar flow to the first The relationship between continuous units in a layer. If have There are 1 direct preceding element, all located in the 1st position. Layers. Multiple pre-flow merging:

[0054] .

[0055] Continuity constraints (ignoring unit switching operation time), for example It is the first Layer Front unit The completion time of the subsequent units. The start time is:

[0056] ,

[0057] Each upstream front unit Compared to The waiting time (used to quantify "kiting wait") is the kitting wait time for a single front unit:

[0058] .

[0059] This implementation agreement: [Shareholding] RSPPT unit Column vector The transpose is denoted as (Row vector); the starting time vector to be determined End time vector The earliest feasible solutions for both are determined by matrix operation constraints.

[0060] After defining the unit parameters, construct a diagonal duration matrix based on the actual duration of each unit. The expression is In the matrix, the diagonal elements represent the actual duration of the corresponding unit, and the off-diagonal elements are... Construct a structure matrix based on the dependencies between units. The matrix element rules are as follows:

[0061] .

[0062] Construct external control vectors based on external constraints such as resource calendar, approval thresholds, and device unlocking times. , The rules for vector elements are:

[0063] .

[0064] The time vector solution is based on the "maximum-additive algebra", and the addition operation is defined. To obtain the maximum value ( Multiplication operation For ordinary addition ( The plan constraints are satisfied: External control constraints are met The time definition satisfies Constraints will be merged into , combined The earliest feasible start and completion times for each unit can be obtained by iterative solution.

[0065] For multiple pre-flow merging scenarios, according to the formula Calculate the kitting wait time for a single edge using the formula. Calculate the global total settling wait time, where the summation range covers all unit pairs with dependencies.

[0066] During production execution, real-time monitoring is conducted to detect changes in the actual duration of units or external control constraints. When a parameter change is detected, the affected set, consisting of the affected units and their downstream reachable units, is determined. The time parameters of each unit are recalculated only within this set according to topological order. The calculation formula is as follows: , ,in For unit The direct predecessor set eliminates the need to recalculate all network elements, enabling rapid dynamic correction of the planning scheme.

[0067] Meanwhile, based on the assembly waiting time, the critical path of the production design is identified, and the project duration and waiting time are optimized by compressing the duration of critical units, adjusting external control, and reconstructing unit dependencies to adapt to the planning needs of different production scenarios.

[0068] Example:

[0069] This example, based on the completed RSOP unit configuration, only fully verifies the time calculation, kitting wait, closed-loop adjustment and optimization scheme of the RSOPT model, without going into detail about the basic resource configuration of the RSOP unit.

[0070] First, we select the basic scenario of dual-front-flow merging. The layer is configured with two front units. and ,in:

[0071] The time parameter is The first front unit of the i-th layer was completed in 63 minutes.

[0072] The time parameter is The second front unit of the i-th layer was completed in 50 minutes.

[0073] According to the multi-preceding merging time propagation rules, subsequent units The earliest feasible start date is:

[0074] .

[0075] Calculate the assembly waiting time of the upstream unit:

[0076] The corresponding waiting time is ,

[0077] The corresponding waiting time is .

[0078] Setting unit Baseline duration The intra-cell delay term is not considered for the time being. Then unit The completion time is .

[0079] Further build a four-unit production chain, including CNC machining unit Sheet metal forming unit Assembly flow unit, The actual time for each inspection unit is calculated as follows:

[0080] Baseline duration Disturbance term Actual duration ;

[0081] Baseline duration Disturbance term Actual duration ;

[0082] Baseline duration Disturbance term Actual duration ;

[0083] Baseline duration Disturbance term Actual duration .

[0084] Construct a diagonal duration matrix based on the above parameters. ; Structure matrix

[0085] External control vector .

[0086] Iteratively compute the time vector in topological order:

[0087] first step: , ;

[0088] Step Two: ;

[0089] Step 3: , , .

[0090] Final start time vector End time vector Waiting time is calculated side by side. for , for , for Global total waiting time .

[0091] Conduct closed-loop adjustment verification: when The actual duration is determined by Increase to When, the affected set is Local recalculation of , of , , of Total waiting time increased to ;

[0092] when The actual duration is determined by Down to At that time, the affected set is After recalculation of , The start date remains unchanged. The total waiting time becomes Terminal completion time remains constant;

[0093] when External control Depend on Change to When, recalculated of , The waiting time for all edges is Total waiting time The terminal completion time is still [time]. .

[0094] Conduct optimization scheme verification:

[0095] Option A (Adjusting Duration to Compress Key Upstream Elements): ... Actual duration optimized to Construct a new diagonal duration matrix Structure matrix External control vector Keeping it unchanged, recalculate to get , all Completed start Completed , Completed The waiting time for all edges is Total waiting time The terminal completion time is shortened compared to the original plan. .

[0096] Option B (Adjusting external control to delay non-critical upstream): Constructing a new external control vector diagonal duration matrix With structure matrix Keeping it unchanged, recalculate to get start Completed , start Completed , Completed The waiting time for all edges is Total waiting time The terminal completion time remains unchanged.

[0097] Option C (Adjusting the structure to split and merge the flow units): Maintain Total duration Unchanged, split it into (duration) )and (duration) Adjust the dependency to , , , Construct a new diagonal duration matrix New structure matrix Given a 5×5 matrix, there are dependent position labels. Location-independent External control vector Keep Unchanged. Recalculated. Completed , Completed , start Completed , start Completed , Completed The waiting time for each side is only for Total waiting time Terminal completion time reduced to .

[0098] The three optimization schemes can be selected or combined according to the actual production needs of schedule, waiting time and resource allocation, fully verifying that the method can realize dynamic planning, time optimization and adaptive adjustment of complex production design, and fully supporting all the technical solutions of the present invention.

[0099] The above description is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several improvements without departing from the principle of the present invention, and these improvements should also be considered within the scope of protection of the present invention.

Claims

1. A complex production design and planning method based on the RSOPT model, wherein the RSOPT model represents a production design unit by a quintuple consisting of resources, site, operation, product, and time parameters, wherein the time parameters include start time, end time, and kitting wait time, characterized in that: The method includes the following steps: The production design process is decomposed into several five-tuple units, and the resources, site, operation, product and process duration of each unit are determined. A structural matrix is ​​established based on the inter-unit dependencies, a diagonal duration matrix is ​​established based on the process duration, and an external gate control vector is established based on external time constraints. Based on the maximum-addition algebra, the start and end times of each unit are calculated according to the structure matrix, diagonal duration matrix and external gate control vector; For units with multiple preceding inputs, the kitting waiting time is calculated based on the start time and the end time of each preceding unit.

2. The complex production design and planning method based on the RSOPT model as described in claim 1, characterized in that: The process duration includes a baseline duration and a disturbance term. The baseline duration is determined based on the process specifications and historical data, while the disturbance term is determined based on time series prediction or real-time monitoring.

3. The complex production design and planning method based on the RSOPT model as described in claim 1, characterized in that: The resources include raw materials, work-in-process, machine time, labor time, fixture space, energy quota, and buffer space; The site may be a physical site, a virtual site, or a hybrid site; The operation includes steps, processes, or processing methods; The products include intermediate or final results.

4. The complex production design and planning method based on the RSOPT model as described in claim 1, characterized in that: Determine the resources, site, operations, products, and process durations for each unit, including: Obtain historical project structure data, resource inventory logs, process specifications, and Gantt chart time logs from CAD, PLM, ERP, or WMS systems; The quintuple unit is generated based on the data.

5. The complex production design and planning method based on the RSOPT model as described in claim 1, characterized in that: In the structure matrix, the corresponding element takes a value of zero when there is a dependency relationship between units, and takes a value of negative infinity when there is no dependency relationship.

6. The complex production design and planning method based on the RSOPT model as described in claim 1, characterized in that: Maximum-addition algebraic computation includes: The maximum value operation is used as the addition operation, and ordinary addition is used as the multiplication operation; The starting time vector is determined based on the result of the multiplication of the structure matrix and the end time vector and the result of the addition of the external gate control vector. The end time vector is determined based on the result of the multiplication operation between the diagonal duration matrix and the start time vector.

7. The complex production design and planning method based on the RSOPT model as described in claim 1, characterized in that: The calculation of the kitting wait time for units with multiple preceding inputs includes: For each direct preceding unit, calculate the current unit's start time and the preceding unit's end time according to the formula: Calculate the waiting time corresponding to this pre-processor unit, where This represents the waiting time from the k-th preceding unit in the i-th layer to the current unit. Indicates the start time of the current cell. This represents the end time of the k-th pre-unit in the i-th layer; The kitting wait time of this unit is obtained by summing the waiting times of all directly preceding units.

8. The complex production design and planning method based on the RSOPT model as described in claim 1 or 7, characterized in that: Also includes: Based on the dependencies between units, according to the formula Calculate the total kitting wait time, where Indicates the total waiting time for all items to be completed. Indicates the start time of the downstream unit. This indicates the end time of the upstream unit, and the summation range covers all unit pairs with dependencies.

9. The complex production design and planning method based on the RSOPT model as described in claim 1, characterized in that: It also includes: monitoring changes in the process duration or external time constraints; determining the affected set consisting of the affected units and their downstream reachable units; and recalculating the start time, end time, and kitting wait time only within the affected set in topological order.

10. The complex production design and planning method based on the RSOPT model as described in claim 1, characterized in that: It also includes: identifying critical paths based on the kitting wait time; By adjusting the matching relationships of resources, site, operation, or product inputs and outputs to compress the process time of units on the critical path, or by adjusting external time constraints to delay the start time of units on the non-critical path, or by adjusting the dependencies between units to restructure the production process, the kitting wait time or end-of-life completion time can be optimized.

Images