Parallel Computing, Multiobjective Optimization and the Workstation CACSD Environment

Abstract In many control system design exercises which employ optimization, the bulk of the computational effort is devoted to the evaluation of the objectives of the optimization at each iteration, This paper demonstrates how, using a new gateway, parallel processing can be used Within the framework of existing computer aided control system design tools to compute these objective values. In the first instance, a simple mapping of objectives to processors is employed. In the likely event that this yields a poorly balanced processor utilisation, this strategy is extended upon by using parallel processing in the evaluation of individual ohjective functions.