Thomas Lacksonen
Dept. of Industrial and Manufacturing Systems Engr.
270 Stocker Center
Ohio University
Athens, Ohio 45701-2979
Abstract
Discrete-event simulation is a significant analysis tool for designing complex systems. Several deterministic search algorithms have been linked with simulation to solve real-world problems, but there are few empirical comparisons of the various algorithms. This paper compares the Hooke-Jeeves pattern search, Nelder-Mead simplex, simulated annealing, and genetic algorithm optimization algorithms on variations of four realistic-sized simulation models. The simulation models include combinations of real variables, integer variables, non-numeric variables, deterministic constraints, and stochastic constraints. The genetic algorithm was the most robust, as it found near best solutions for all twenty-five test problems. However, it required the most replications of all the algorithms. The pattern search algorithm also found near best solutions to small and medium sized problems with no non-numeric variables, while requiring fewer replications than the genetic algorithm.
submitted to Computers and Industrial Engineering 1997