Abstract

multi-objective optimization on complex stochastic systems. A stochastic simulation model estimates the results of parameter settings for the system. A fuzzy ordinal preference model aggregates these results into a single fitness value for the input parameter set. A genetic algorithm uses this fitness value to search for a population of high performance parameter sets which achieve the modeler’s objectives. In a tactical planning experiment, Fuzzy-Genetic Decision Optimization enabled a human planner to develop a significantly better plan than he developed without automated assistance. 1

Citations