Vehicle Assembly Sequence Prediction for Workstation Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Designing a manufacturing process for vehicles is time-consuming due to the complexity of allocating parts, tools, and steps across multiple workstations, with unclear resource allocation for new or existing processes, especially when precedence and clearance issues are involved.
Innovation Solution
A method is developed to predict the sequence of manufacturing processes by retrieving surrogate process segments from a historical database, using similarity analysis and a sequence inference model to generate a target process that defines process steps and workstation allocation, based on textual and sequence similarities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If engineers manually determine resource allocation for manufacturing processes, then they can handle complex precedence and clearance issues, but the process becomes time-consuming
Solution Approach 1:
The system performs preliminary analysis by retrieving historical process segments and pre-calculating similarity metrics before final process design. This allows engineers to leverage pre-processed historical data rather than starting from scratch, significantly reducing design time while maintaining allocation correctness through systematic comparison of process segments based on textual and sequence similarities.
Solution Approach 2:
The system copies and adapts proven process segments from historical data to new manufacturing processes. By retrieving surrogate process segments that are textually and sequentially similar to target processes, the system replicates successful resource allocation patterns from history, ensuring correctness while eliminating redundant manual analysis time.
2Productivity
If historical data is used to determine resource allocation, then the process becomes faster, but the allocation of resources for new or existing processes remains uncertain
Solution Approach 1:
The system implements feedback by systematically comparing historical process segments with target processes using similarity metrics. The sequence inference model uses this feedback to iteratively refine resource allocation predictions, cross-validating against multiple historical examples to ensure both speed and certainty in the allocation process.
Solution Approach 2:
The system changes parameters by computing multiple similarity metrics (textual similarity, sequence similarity) and using these as inputs to the sequence inference model. This multi-parameter approach transforms uncertain historical data into reliable predictions by weighing different aspects of process segment comparison to determine optimal resource allocation.
3Measurement precision
If multiple similarity metrics are computed for process segments, then the accuracy of sequence prediction improves, but the computational complexity increases
Solution Approach 1:
The system segments the similarity assessment into distinct components: textual similarity measurement and sequence similarity measurement. Each component handles a specific aspect of comparison independently, then their results are combined by the sequence inference model. This segmentation improves accuracy through comprehensive comparison while managing complexity by dividing the analysis into manageable, specialized sub-tasks.
Data Source
AI summary
A method of predicting sequence for a manufacturing process to assemble a product includes obtaining a target plan for the manufacturing process. The target plan defines a process segment to be performed during the manufacturing process. The method includes retrieving a plurality of surrogate process segments based on the target plan, determining, for each surrogate process segment from the plurality of surrogate process segments, a segment similarity value based on at least one of a textual similarity and a sequence similarity between the surrogate process segment and the process segment of the target plan, and generating a target process defining a process sequence for performing the manufacturing process based on the segment similarity values of the plurality of surrogate process segments and a sequence inference model.