Multi-objective optimisation of product modularity

The optimal modular configuration of a product’s architecture can lead to many advantages throughout the product lifecycle. Advantages such as: ease of product upgrade, maintenance, repair and disposal, increased product variety and greater product development speed. However, finding an optimal modular configuration is often difficult. Finding a solution will invariably mean trade-offs will have to be made between various lifecycle drivers. One of the main strengths of a computerised optimisation is that trade-off analysis becomes simple and straightforward and hence speeds up the product architecture decision making process. However, there are a lack of computerised methods that can be applied to optimise modularity for multiple lifecycle objectives. To this end, an integrated optimisation framework has been developed to optimise modularity from a whole lifecycle perspective, namely, design, production, use and end of life. For each lifecycle phase there are two modularity criteria- module independence and module coherence. The criteria that fall under the category of module independence evaluate the degree of coupling between the products components, coupling can be physical, functional or design based. Criteria under module coherence, evaluate the similarity of modular drivers between components. The paper will examine the developed optimisation framework and software prototype. The prototype software uses a number of matrixes to represent the product architecture. A goal based genetic algorithm is used to search the matrixes for modular configurations that most satisfies the criteria of the four lifecycle phases. Sensitively analysis is carried out by changing the goal weights.