Multiagent based state transition algorithm for global optimization

TL;DR

This paper introduces MASTA, a multiagent state transition optimization algorithm with linear convergence, which effectively finds optimal solutions through iterative population updates and communication, outperforming existing algorithms on benchmark functions.

Contribution

The paper proposes a novel multiagent based state transition algorithm with linear convergence rate, enhancing global optimization performance.

Findings

MASTA demonstrates superior performance on benchmark functions.

The algorithm achieves linear convergence rate.

It outperforms some state-of-the-art optimization algorithms.

Abstract

In this paper, a novel multiagent based state transition optimization algorithm with linear convergence rate named MASTA is constructed. It first generates an initial population randomly and uniformly. Then, it applies the basic state transition algorithm (STA) to the population and generates a new population. After that, It computes the fitness values of all individuals and finds the best individuals in the new population. Moreover, it performs an effective communication operation and updates the population. With the above iterative process, the best optimal solution is found out. Experimental results based on some common benchmark functions and comparison with some stat-of-the-art optimization algorithms, the proposed MASTA algorithm has shown very superior and comparable performance.