Hybrid parameter control approach applied to a diversity-based multi-objective memetic algorithm for frequency assignment problems

Abstract

In order to address the difficult issue of parameter setting within a diversity-based Multi-objective Evolutionary Algorithm (MOEA), we recently proposed a hybrid control scheme based on both Fuzzy Logic Controllers (FLCs) and Hyperheuristics (HHs). The method simultaneously adapts both symbolic and numeric parameters and was shown to be effective when controlling a diversity-based MOEA applied to a range of benchmark problems. Here, we show that the hybrid control scheme generalises to other meta-heuristics by using it to adapt several parameters of a diversity-based multi-objective Memetic Algorithm (MA) applied to a Frequency Assignment Problem (FAP). Using real-world instances of the FAP, we demonstrate that our proposed parameter control method outperforms parameter tuning of the MA. The results provide new evidence that the method can be successfully applied to significantly more complex problems than the benchmarks previously tested.

Towards Optimal Frequency Plans: A Survey of Frequency Assignment Strategies, Models, and Methods

“…The strength of multi-objective optimization has motivated researchers to develop a multi-objectivized version of the FAP [49]. In this form, the problem is treated as a single-objective minimization task in a multi-objective framework with a main objective and additional objectives that measure diversity in terms of the distance to the closest neighbour [50] or the average distance to all individuals, and which may be maximized [51][52][53].…”

“…One generalization assumes that, given a simple graph 𝐺 = (𝑉, 𝐸), each node 𝑖 of 𝑉 is associated with a list 𝐿 of allowed colours [51,52]. The lists are not necessarily of the same length.…”