Production Management Support Apparatus for Worker Skill Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In high-mix, low-volume production environments, estimating standard work time and failure occurrence rate for workers with varying skill levels and no work history is challenging, leading to discrepancies in production scheduling and worker allocation.
Innovation Solution
A production management support apparatus that acquires capability and work performance information to generate predictive models for estimating work status, allowing for accurate allocation of workers to tasks based on their skills and past performance data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard work time is determined based on actual measurement using stopwatch, then measurement precision is improved, but loss of time increases due to requiring measurements of all workers and new measurements for every new worker or new work
Solution Approach 1:
The system performs preliminary measurements and calculations during the measurement phase, storing the results in advance. When new workers or new work types are introduced, the pre-established relationships and data can be quickly applied without requiring complete re-measurement, thus reducing the time loss while maintaining measurement precision.
Solution Approach 2:
The system creates a standardized work time model based on measured data that can be copied and applied to similar workers and work types. Instead of measuring every worker individually, the system uses representative measurements that can be replicated across similar cases, significantly reducing measurement time while preserving accuracy through the predictive model.
2Loss of time
If standard work time is calculated based on past work history, then loss of time is reduced, but measurement precision deteriorates when workers have no experience and therefore no work history
Solution Approach 1:
The predictive model serves multiple functions: it can estimate standard work times for workers with history, workers without history, and can adapt to new work types. By training the model on aggregated historical data from multiple workers, it creates a universal estimation capability that works across different scenarios, including cases where individual worker history is unavailable.
Solution Approach 2:
The predictive model acts as an intermediary between historical data and individual worker estimation. Instead of directly using past work history (which may be incomplete for new workers), the model processes and synthesizes historical patterns to generate accurate estimates for individual workers, even those without personal work history, by leveraging organizational-level historical data.
3Device complexity
If uniform standard work time is determined without taking worker characteristics into consideration, then device complexity is reduced, but manufacturing precision deteriorates due to discrepancy from actual work time
Solution Approach 1:
The system transitions from static uniform standard work times to dynamic personalized estimates. The predictive model automatically adjusts standard work time estimates based on individual worker characteristics (skill levels, experience, performance history) without requiring manual intervention. This dynamic adaptation maintains high manufacturing precision while the automated nature of the model keeps system complexity manageable.
Data Source
AI summary
A CPU 501 of a production management support apparatus 500 is configured to execute: a process of acquiring capability information of a plurality of workers; a process of acquiring work information of a plurality of kinds of work; a process of acquiring work performance information including a work status of a plurality of kinds of work actually executed by the plurality of workers; a process of accepting a designation of a target worker and target work; a process of acquiring the work performance information of the plurality of workers related to the plurality of kinds of work, and generating a predictive model which adopts a work status in the work performance information of each work as a target variable, and which adopts any plurality of feature amounts among feature amounts included in the work information and the capability information as an explanatory variable; a process of applying the predictive model to the capability information of the target worker and the work information of the target work, to estimate a work status when the target work is performed by the target worker; and a process of displaying the estimated work status.