PlumX Metrics
Embed PlumX Metrics

Multi-space evolutionary search with dynamic resource allocation strategy for large-scale optimization

Neural Computing and Applications, ISSN: 1433-3058, Vol: 34, Issue: 10, Page: 7673-7689
2022
  • 4
    Citations
  • 0
    Usage
  • 7
    Captures
  • 0
    Mentions
  • 0
    Social Media
Metric Options:   Counts1 Year3 Year

Metrics Details

Article Description

Multi-space evolutionary search (MSES) is a recently proposed paradigm to optimize across multiple solution spaces for solving a large-scale optimization problem. MSES allocates the computational resource equally on each search space. However, different search spaces are likely to make different contributions to the finding of global optimum of the given problem. Dividing the limited resources equally on each search space in MSES is thus not an efficient strategy. This paper aims to investigate how to utilize the imbalanced efforts to allocate computational resources in multiple search spaces to efficiently improve the performance of MSES. In this paper, we propose a novel multi-space evolutionary search with dynamic resource allocation strategy (MSES - DRA) for large-scale optimization. In particular, a detection mechanism is presented to measure the reasonableness of assignment in terms of computation resource of different spaces. Further, according to the interaction between optimal individual and population, the proposed dynamic resource allocation strategy is designed based on the explicit–implicit contributions of spaces. The explicit and implicit contributions are defined by the fitness improvement of best solution and the survival of individuals, respectively. An adaptive technology based on the feedback is conducted to balance the assignment of computational resources for each search space. To evaluate the performance of the proposed method, comprehensive empirical experiments have been conducted on the CEC2013 large-scale benchmark problems.

Bibliographic Details

Provide Feedback

Have ideas for a new metric? Would you like to see something else here?Let us know