Affiliations: [a] REGIM-Lab, National Engineering School of Sfax, University of Sfax, Sfax, Tunisia | [b] ISSAT Sousse, University of Sousse, Sousse, Tunisia | [c] Machine Intelligence Research Labs, Scientific Network for Innovation and Research Excellence, Auburn, WA 98071, USA | [d] FST, Hassan 1st University, Settat, Morocco
Correspondence:
[*]
Corresponding author: Nizar Rokbani, REGIM-Lab, National Engineering School of Sfax, University of Sfax, Sfax, Tunisia. E-mail: Nizar.rokbani@ieee.org.
Abstract: Combinatorial optimization problems have several industrial applications such as Network routing, IOT network routing, path Planning for robotics and manufacturing for which the travelling Salesman Problem, TSP, can serve as typical test bench. This paper investigates new variants of the Fuzzy Ant Supervised by PSO, FAS-PSO and Simplified Ant Supervised by PSO, SAS-PSO coupled with a local search, Ls, mechanism. The proposed method is based on the Fuzzy PSO to supervise and tune ACO parameters, in addition to a local search mechanism helping in avoiding cities local crossing. The SAS-PSO-Ls uses the same idea while with the simplified PSO as supervisor. Experimentations (a space is missed before “Experimentations”) and results are based TSP test benches with a statistical analysis and a comparative study with the standard AS-PSO and similar state of art methods. FAS-PSO-Ls gives better than the state of art for eil51, berlin52, while the SAS-PSO-Ls is giving better results for the following cases: eil51, berlin52, st70.
