A global method for the limited K-partitioning of hypergraphs representing optimal design problems in complex machine systems

Purpose - The purpose of this paper is to find a global method for the limited K-partitioning of hypergraphs representing optimal design problems in complex machine systems. Design/methodology/approach - To represent some real design considerations, a new concept of semi-free hypergraphs is proposed and a method to apply semi-free hypergraphs to the decomposition of complex design problems based on optimal models is also suggested. On this basis, the limited K-partitioning problem of semi-free hypergraphs and its partitioning objective for the optimal design of complex machines is presented. A global method based on genetic algorithms, GALKP, for the limited K-partitioning of semi-free hypergraphs is also proposed. Finally, a case study is presented in detail. Findings - Semi-free hypergraphs are a more powerful tool to map a complex engineering design problem. The decomposition of complex design problems may be converted to a limited K-partitioning problem of semi-free hypergraphs. The algorithm presented in this paper for the limited K-partitioning of semi-free hypergraphs is fast, effective, and powerful. Research limitations/implications - The traditional methods based on hypergraphs have some limitations while applied to the decomposition of some complex problems such as the design of large-scale machine systems. The proposed method is helpful to solve similar engineering design problems. Practical implications - The paper illustrates a faster and more effective method to implement the decomposition of large-scale optimal design problems in complex machine systems. Originality/value - This paper shows a new way to solve the complex engineering design problems based on semi-free hypergraphs and its K-partitioning method.