Using metaheuristics to develop production wheels
When it comes to production planning, the need for balancing flexibility and stability is crucial in a wide range of industries. On the one hand, companies strive to remain flexible in order to match customer demand. On the other hand, there is a need for the leveling of production to create stability and operate efficiently. Often, manufacturers apply batch logic for their planning. It is, however, possible to step towards leveled production and to produce in line with customer demand at the same time. This simultaneously creates the stability essential in driving a virtuous circle of standard work, continuous improvement and increased performance. Production wheels can be used to realize this.
A production wheel is a method to schedule the production in a regularly repeating sequence. By integrating demand over a certain period of time and leveling production in this time period, a production wheel aims to balance the need for flexibility and the need for efficiency.
Two-step method for designing optimal production wheels
Existing methods for designing production wheels are often suboptimal and rely on manual judgement. The method developed by EyeOn, however, applies analytical models for the design and optimization of production wheels. The method consists of the following two steps:
- Allocate all products to dedicated production lines or machines based on changeover costs. A ‘randomized variable neighborhood search’ is applied to find a good solution to this problem.
- Develop a repeating, cyclical production plan for each production line or machine. By balancing inventory costs and changeover costs, a genetic algorithm finds high-quality production wheels.
Case study: Applying the EyeOn production wheel method in a large multinational (process industry)
We performed a case study to test the proposed methodology for designing production wheels. In a production setting with multiple machines and a broad product mix, we proved that the proposed models perform well in designing a cost-efficient production wheel. Our method decreased the quarterly costs by 10.0% compared to an earlier customized production wheel developed by EyeOn and by 39.5% in comparison with the existing production plan of the company in question.
Balancing inventory and changeover costs is key in designing production wheels. By decreasing batch sizes, the proposed method finds a better balance than the benchmark production plans. A production plan with smaller batches and more changeovers leads to higher costs of changeovers but greatly reduces inventory costs (see figure). In the end, this results in significantly lower total costs.
Conclusion
Production wheels enable you to create stability and regularity while remaining flexible towards your customers. But developing production wheels can be a complex task. Simple heuristics with manual judgements are only a first step towards a high-quality production plan. Therefore we developed this new analytical approach using optimization algorithms to best support you in developing production wheels.
For more information feel free to contact us or reach out to Dan Roozemond directly!