NPS investigates new methods for priority-rule-based scheduling and control of flexible, volatile production processes at the operational shop floor level according to user-defined strategies. To do so, complex priority rules are automatically synthesized and iteratively optimised by using meta-heuristics and simulation models. To specifically define the desired production strategies, economic, environmental and work psychological indicators are merged by holistic rating models.
Background / Problem: The competitiveness of manufacturing companies increasingly depends on the quality of the operational production planning and control. Due to an increasing volatility (e.g. not predictable, stochastic influences; varying boundary conditions) forward planning must be substituted more and more by situational decisions (production control by dispatching). At the same time, an increasing pressure is observed, not only to exploit existing means of production and resources efficiently, but also to design and operate the production processes in accordance with aspects of sustainability but without economic losses.
Objectives / Methods: The objective is to use complex priority rules for the situational decision support of the operational guidance and control of volatile production processes on shop floor level defined by strategies. For this purpose meta-heuristic optimisation methods are combined with simulation-based logistical process and rating models to automatically synthesize dispatching rules and continually optimize them in terms of a closed control loop. With the aim of a holistic, multi-objective optimization, economical (e.g. throughput, utilization, stocks, due date deviation), environmental (e. g. energy consumption, emissions, noise emissions) and work psychological indicators (e. g. stress, monotony) are merged together by rating models. More specifically it will be investigated how well dispatching performs in the case that many constraints (e. g. availability, needs, relationships, costs) vary dynamically and stochastically.
As use cases for the evaluation and for the iterative improvement of the methods and procedures developed a highly flexible production of casting plastic parts as well as the production of bended and stamped parts will serve.
Results / Findings: The methods and techniques to be investigated and developed will represent new approaches for advanced planning and scheduling (APS) and shall be incorporated in future, sustainable Manufacturing Execution Systems (MES). Besides the findings on the applicability and robustness of the method, the project will deliver software components and simulation libraries, as well as guidelines and algorithms for the automatic generation and optimization of complex priority rules as results.
Scheduling, production control, holistic rating, priority rules, genetic programming, rule synthesis, optimisation
Development of methods and processes for sustainable acting decision support systems to be used in operative production planning and Control
Produktion der Zukunft