Surrogate-Assisted Particle Swarm with Local Search for Expensive Constrained Optimization

This paper develops a surrogate-assisted particle swarm optimization framework for expensive constrained optimization called CONOPUS (CONstrained Optimization by Particle swarm Using Surrogates). In each iteration, CONOPUS considers multiple trial positions for each particle in the swarm and uses surrogate models for the objective and constraint functions to identify the most promising trial position where the expensive functions are evaluated. Moreover, the current overall best position is refined by finding the minimum of the surrogate of the objective function within a neighborhood of that position and subject to surrogate inequality constraints with a small margin and with a distance requirement from all previously evaluated positions. CONOPUS is implemented using radial basis function (RBF) surrogates and the resulting algorithm compares favorably to alternative methods on 12 benchmark problems and on a large-scale application from the auto industry with 124 decision variables and 68 inequality constraints.