Model Implementation

In the absence of a tractable mathematical structure of Problem 13.1, we study the behavior of heuristics, “quick and dirty” algorithms which return feasible solutions that are not necessarily optimal. In particular, we apply a methodology called Simulation/Optimization which is a general expression for solving problems where one has to search for the settings of controllable decision variables that yield the maximum or minimum expected performance of a stochastic system as presented by a simulation model [146, 148]. For a compact picture of the fieldfield, see the reviews of the Winter Simulation Conference [147, 149]. Given the analogies between the TSP and the PTSP, it is reasonable to expect that, like in the TSP, a good heuristic for the problem may be obtained by the integration of 1. A solution construction algorithm generating candidate solutions 2. A local search algorithm, which tries to improve as much as possible the candidate solution The sequence construction and improvement of a solution is repeated several times until a good solution or some other termination criterion is not satisfied. With respect to Fig. 14.1, this involves running a simulation for an initial set of values, analyzing the results, changing one or more values, rerunning the simulation, and repeating the process until a satisfactory (optimal) solution is found ([123], p. 13).