Abstract
This study develops the system which assists the unit placement decision of the automatic assembling equipments to assemble efficiently. The unit is a part supply or a robot hand constituting assembling equipment. Conventional research has used reinforcement learning to determine efficient unit placement but this research adopts Genetic Algorithm (GA) instead of Reinforcement Learning. First, as input data, information such as an arrangeable region of each unit is input, and the input region is divided into fixed sizes. Next, the unit is placed in the divided region. Prepare a group of individuals whose genes are the placement of each unit, and repeat selection, crossover, mutation, and evaluation to improve placement. This system introduces the concept of Big Mutation that changes the mutation rate of GA every certain generation. As a result, the possibility of searching for an optimal unit placement increased, and it is expected that the system simulation time is shortened.
