Multicriteria decision making with parallel clusters in structural topology optimization

Generating Pareto surfaces is a well accepted technique in multiattribute decision making. Alternatively, an interactive approach may be used where designs on the Pareto frontier are compared or rank ordered and newer designs are generated to maximize an implicit utility function. When each function evaluation involves a finite element analysis or other type of simulation, multicriteria approaches can become highly computationally expensive. In this context, techniques for reducing computing time using parallel processors are discussed. A structural topology problem is chosen as a test problem and Pareto surfaces are generated using two different clusters. Near linear speed-ups and high efficiencies are observed on both the clusters. It is possible to integrate this methodology into computer programs that assist design engineers in the industry leading so that less turn around time is required to make critical decisions.