A Unifying Survey of Reinforced, Sensitive and Stigmergic Agent-Based Approaches for E-GTSP

Abstract: The Generalized Traveling Salesman Problem (GTSP) is one of the NP-hard combinatorial optimization problems. A variant of GTSP is E-GTSP where E, meaning equality, has the constraint: exactly one node from a cluster of a graph partition is visited. The main objective of the E-GTSP is to find a minimum cost tour passing through exactly one node from each cluster of an undirected graph. Agent-based approaches involving are successfully used nowadays for solving real life complex problems. The aim of the current paper is to illustrate some variants of agent-based algorithms including ant-based models with specific properties for solving E-GTSP.