ANALYTIC HIERARCHY PROCESS AND ARTIFICIAL NEURAL NETWORKS MODEL FOR MANAGEMENT INFORMATION SYSTEMS SOFTWARE SELECTION IN COMPANIES

Abstract

Making a decision is one of the most important activities in business life. Managers need correct and reliable estimations for this. Considering scientific criterions provide better results while decision making. Decision problem is more generally defined as a making the most appropriate choice in alternatives set according to at least one aim or one criterion. Analytic Hierarchical Process (AHP) is one of the most widespread techniques for selection of the most appropriate alternative. Alternatives weights can be determined qualitatively by setting multi-level decision structures and forming pair comparison matrixes according to decision makers' subjective judgments. In recent years, another subject that has wide implementations is Artificial Neural Networks (ANN). In decision making process, network can be trained through supervised learning according to firms' tendency in sector. Thus differences between alternatives can be brought up without determining importance of criterions by calculating alternative scores using Multi Layer ANN. In this article, AHP and ANN methods are implemented in software selection problem with nine criterions and five alternatives using Expert Choice and NeuroSolutions programs and analyzed results are compared. Besides in this study, Multi Layer ANN solutions at different topologies are evaluated for selection problem.