SEARCH ALGORITHMS APPLIED TO PARAMETERS ESTIMATION IN AN INVENTORY MANAGEMENT PROBABILISTIC MODEL

Abstract

Inventory management is an important instrument regarding to the efficiency of modern organizations. Inventory models demand parameters estimation, which are usually associated to uncertainties. Such parameters are generally estimated by mathematical modeling. However, in some real situations the models may be distant from reality owing to the input data suppositions undertaken. Besides, the search for the optimal solution in the whole solution domain is infeasible in most cases. Heuristics search algorithms are an alternative to solve problems of this nature. In this work, Simulated Annealing, Tabu Searching, Hill Climbing and Genetic Algorithm techniques were applied and tested in a particular inventory management problem. Results were compared to those obtained by exhaustive technique. The proposed algorithms were implemented and tested successfully and the results are in agreement with those obtained by exhaustive searching algorithm, thus justifying these algorithms as an alternative for problems of this nature.