Simulation optimization with search techniques: finding optimal combination of kanbans

When the systems under investigation are complex and inherit uncertainties, the analytical solutions to these systems become impossible. Because of the complex stochastic characteristics of such systems, simulation can be used as an analysis tool to predict the performance of an existing system or a design tool to test new systems under varying circumstances. Simulation model in which the set of output is estimated for given a particular set of input is an input/output model. Since simulation is not an optimization tool, the difficulty is then to know how to drive simulation experiments in order to determine the value of decision variables. In fact, when there is large set of decision variables with many possible values, the number of possible combinations is such that an exhaustive search is not possible. Recently, search techniques have been used with simulation model to overcome this difficulty. In this study, Tabu Search (TS), Genetic Algorithms (GA) and Simulated Annealing (SA) belonging to general purpose search techniques are applied to find optimum combination of kanbans on a hypothetic Just in Time production system stemmed from a real production system of mobile phone. The effectiveness and efficiency of the general purpose search algorithms together with random search are investigated according to solution quality and solution time.