Recurring two-stage evolutionary programming: a novel approach for numeric optimization.

Area of Science:

• Computational intelligence
• Optimization algorithms
• Evolutionary computation

Background:

• Balancing global exploration and local exploitation is crucial for evolutionary algorithms (EAs) in complex problem solving.
• Conventional EAs often struggle to maintain this balance, impacting their ability to find near-optimum solutions.

Purpose of the Study:

• To introduce a novel approach, Recurring Two-Stage Evolutionary Programming (RTEP), for balancing exploration and exploitation in EAs.
• To analyze the necessity and effectiveness of alternating exploration and exploitation stages.

Main Methods:

• RTEP employs a repeated and alternated execution of two distinct stages: exploration and exploitation.
• Each stage features unique mutation operators, selection strategies, and objectives tailored for exploration or exploitation.
• Analytical and empirical studies were conducted, including tests on 48 benchmark numerical optimization problems.

Main Results:

• The experimental results demonstrate the significant effectiveness of RTEP.
• The repeated and alternated exploration and exploitation operations employed by RTEP yield superior performance.

Conclusions:

• RTEP provides an effective mechanism for balancing exploration and exploitation in evolutionary algorithms.
• The proposed method shows remarkable potential for improving the performance of EAs in complex optimization tasks.