Parallel Social Group Optimization (Psgo): Performance and Its Application to Routing in Wmn’s

Abstract

Abstract This paper proposes a new multi-population-based social optimization technique called Parallel Social Group Optimization (PSGO). The algorithm is inspired by the learning behaviour of humans in different groups. In this algorithm, we consider the complete process of human interaction within the group and outside the group. We implement the proposed algorithm in MATLAB and tested it on 30 standard benchmark functions. For performance analysis purposes, we compare the PSGO algorithm with other recent 16 algorithms. The PSGO algorithm outperformed the other 16 algorithms on 4 standard benchmark functions. None of the other algorithms could match this performance. In addition, for functions f1,f2, f3,f4,f5, and f7of standard benchmark functions it produced the best performance but this performance was equalled by a few other algorithms also. Further, we propose a PSGO-based dynamic route evaluation approach for Wireless Mesh Networks (WMNs). We implemented the PSGO-based routing approach in MATLAB and compared it with 9 soft computing and hard computing-based approaches namely AODV, DSR, ACO, BBO, BAT, Firefly, BBBC, GA, and SGO. On over 1500 dynamic node network situations, the PSGO-based routing technique outscored all other 9 algorithms.