Concept optimization for mechanical product by using ant colony system

The aim of conceptual design is to generate the best design candidate. Concept solving in conceptual design can be viewed as a problem of combinatorial optimization, in which there exists a "combinational explosion" phenomenon when using the traditional morphological matrix method to tackle it. In this research, a concept optimization problem is studied based on an Ant Colony System (ACS). By analyzing the similarity between concept solving and Traveling Salesman Problem (TSP), concept solving is transformed into a problem of optimal path in combinatorial optimization, where the dynamic programming based solution space model and the longest path based optimization model are developed. Then, the ant algorithm to resolve TSP is adopted to implement concept optimization according to the positive feedback searching mechanism of ACS, and some improvements are made incorporating crossover and mutation operators of a genetic algorithm (GA), to obtain the optimal scheme rapidly and effectively. Finally, a conceptual design case of press is given to demonstrate the feasibility and rationality of this proposed approach. The employment of ACS enables concept solving to be implemented with an algorithm and thus possesses better operability, which offers a promising way to solve the "combinatorial explosion" problem in conceptual design.