Analysis of the genetic algorithm Gamic for multicriteria classification of inventories

Gamic (Genetic Algorithm for Multi-criteria Inventory Classification) is analysed in this paper. It is shown how Gamic may produce rank reversal and how, because of this, it is not an appropriate classification method. Moreover, we propose a modification of this algorithm that does not generate rank reversal and does not need all the elements of the inventory in order to carry out the classification. The modification of Gamic that is proposed to solve the former problems consists of the change of the weighted product instead of the aggregation function of the weighted sum, in order to aggregate the priorities in the analytical hierarchy process (AHP), showing how the use of an aggregation function of weighted products in AHP produces both feasible solutions and known ones