An efficient optimisation method based on weighted AND-OR trees for concurrent reconfigurable product design and reconfiguration process planning

A reconfigurable product can be used to replace multiple products through change of its configurations. Since multiple design candidates are usually considered based on design requirements, and multiple reconfiguration process candidates are usually considered to change between two product configurations, a method was developed in our previous research to identify both the optimal reconfigurable product design and its optimal reconfiguration process plan. This method was not effective when the numbers of design candidates and reconfiguration process candidates are large. To solve this problem, an efficient optimisation method is developed in this research based on weighted AND-OR trees for concurrent reconfigurable product design and reconfiguration process planning. First, the nodes in the product design and reconfiguration process AND-OR trees for modelling design and process candidates are associated with weights representing evaluation heuristics. Then a modified beam search method is developed to improve optimisation efficiency by pruning of the nodes in the AND-OR trees that are unlikely to achieve the optimal results. The created solutions modelled by design and reconfiguration process candidates are further ranked by heuristics. The top-ranked solutions are finally evaluated by numerical engineering functions to identify the optimal solution.