Three level hierarchical decision making model with GA. Issue 6 (29th July 2014)

• Record Type: Journal Article
• Authors: D'Amato, Egidio; Daniele, Elia; Mallozzi, Lina; Petrone, Giovanni
• Editors: Massimiliano Vasile, Dr Edmondo Minisci and Dr Domenico Quagliarella, Professor
• Is Part Of: Engineering computations. Volume 31:Issue 6(2014)
• Page Start: 1116
• Page End: 1128
• Publication Date: 2014-07-29
• Subjects: Genetic algorithm -- Hierarchical game -- Stackelberg strategy; Computer-aided engineering -- Periodicals; Computer graphics -- Periodicals; 620.00285
• DOI: 10.1108/EC-03-2012-0075
• Languages: English
• ISSNs: 0264-4401

Abstract

Abstract : Purpose: – The purpose of this paper is to propose a numerical algorithm able to describe the Stackelberg strategy for a multi level hierarchical three-person game via genetic algorithm (GA) evolution process. There is only one player for each hierarchical level: there is an upper level leader (player L0), an intermediate level leader (player L1) who acts as a follower for L0 and as a leader for the lower level player (player F) that is the sole actual follower of this situation. Design/methodology/approach: – The paper presents a computational result via GA approach. The idea of the Stackelberg-GA is to bring together GAs and Stackelberg strategy in order to process a GA to build the Stackelberg strategy. Any player acting as a follower makes his decision at each step of the evolutionary process, playing a simple optimization problem whose solution is supposed to be unique. Findings: – A GA procedure to compute the Stackelberg equilibrium of the three-level hierarchical problem is given. An application to a Authority-Provider-User (APU) model in the context of wireless networks is discussed. The algorithm convergence is illustrated by means of some test cases. Research limitations/implications: – The solution to each level of hierarchy is supposed to be unique. Originality/value: – The paper demonstrates the possibility of using computational procedures based on GAs in hierarchical three level decision problems extending previous results obtained in the classical two level case.