Advanced environmental adaptation method

doi:10.1007/s10489-022-03923-4

Advanced environmental adaptation method

2022-08-05

0

PRE

AI

DOI：10.1007/s10489-022-03923-4

Original

Original request for help

Save

Abstract

En 中文

Many nature-inspired algorithms have been designed to solve optimization problems by combining randomization with exploitation and exploration. Randomization plays a vital role in defining the convergence rate and performance of the algorithm. However, excessive use of randomization may adversely affect the performance of the algorithm. Therefore, a proper balance in randomization, exploitation, and exploration is required to improve the convergence rate of the algorithm. One of the methods to solve the optimization problems is Environmental Adaptation Method. It is a randomized algorithm that works on the theory of adaptive learning. It was followed by an enhanced version named Improved Environmental Adaptation Method. Both of these algorithms used binary encoding to represent the solutions. Since binary encoding requires extra efforts to convert a binary solution to a real solution, a real parameter version of the algorithm will be a good alternative to solve problems. In this paper, we present a new real parameter algorithm named Advanced Environmental Adaptation Method with a novel approach to balance randomization, exploitation, and exploration. This is achieved using operators that make it efficient to search for optimal global solutions. We compare the performance of this new algorithm with other state-of-the-art algorithms. The results show the superiority of the proposed algorithm over existing algorithms. We also demonstrate the effectiveness of the proposed algorithm for real-life problems by applying it to the salient object detection problem, which is an emerging problem in computer vision.

Keywords:

Adaptation

Optimization

Evolutionary algorithm

Salient object detection

Advanced environmental adaptation method

Abstract

Journal

Organization