Abstract
Wireless Sensor Network (WSN) management has several NP-hard optimization problems with multiobjective space and variable solution space nature. Such problems are considered challenging and require a carefully developed and well-understood algorithm in behavior and performance. This article presents an open-source benchmarking framework that enables comparison within supporting multiobjective optimization metrics and networking measures. As the first of its type, the framework is designated as Variable-length Multiobjective Optimization (VLMOO). It includes three sets of mathematical benchmarking problems, WSN deployment problem with multiobjective, and variable-length nature. It is extensible and flexible in enabling interfacing with other algorithms with rich visualization graphs generation for the metrics.
•Proposing the class concept (solutions) under the same dimension.•Converting the search from swarm perspective to class perspective.•Proposing criteria for selecting exemplar within class and for evaluating classes.•Proposing operators for intraclass searching.•Evaluating VLMOPSO functions and WSND problems.