Memory-based adaptive partitioning (MAP) of search space for the enhancement of convergence in Pareto-based multi-objective evolutionary algorithms

• Convergence was better improved by the MAP algorithm than by local search methods.
• MAP-NSGAII improved the convergence accuracy of NSGAII by a factor of 50 in average.
• Memory-based partitioning allowed avoiding unnecessary function calls (40% average).
• Convergence improvements were achieved for both unimodal and multimodal problems.

A new algorithm, dubbed memory-based adaptive partitioning (MAP) of search space, which is intended to provide a better accuracy/speed ratio in the convergence of multi-objective evolutionary algorithms (MOEAs) is presented in this work. This algorithm works by performing an adaptive-probabilistic refinement of the search space, with no aggregation in objective space. This work investigated the integration of MAP within the state-of-the-art fast and elitist non-dominated sorting genetic algorithm (NSGAII). Considerable improvements in convergence were achieved, in terms of both speed and accuracy. Results are provided for several commonly used constrained and unconstrained benchmark problems, and comparisons are made with standalone NSGAII and hybrid NSGAII-efficient local search (eLS).

The flow diagram below illustrates the integration of memory-based adaptive partitioning of search space (MAP algorithm) in the state-of-the-art NSGAII, aiming to improve accuracy/speed ratio in the convergence of combined MAP-NSGAII framework.Figure optionsDownload as PowerPoint slide