A novel multi-hybrid differential evolution algorithm for optimization of frame structures.

Area of Science:

• Computational Intelligence
• Optimization Algorithms

Background:

• Differential Evolution (DE) is a powerful optimization tool for complex problems.
• Existing DE algorithms can be improved for better performance and efficiency.

Purpose of the Study:

• To propose a Multi-Hybrid DE (MHDE) algorithm.
• To enhance the working capability and efficiency of DE without compromising solution quality.

Main Methods:

• Incorporates adaptive parameters, enhanced mutation and crossover, population reduction, iterative division, and Gaussian random sampling.
• Utilizes Weibull distribution and Gaussian random sampling to prevent premature convergence.
• Employs iterative division for improved exploration and exploitation.

Main Results:

• MHDE was validated on IEEE CEC benchmark suites (2005, 2014, 2017).
• Applied to four engineering design problems and three frame design weight minimization problems.
• Outperformed recent hybrid algorithms in statistical tests (Friedman and Wilcoxon rank sum).

Conclusions:

• The proposed MHDE algorithm demonstrates superior performance compared to existing methods.
• MHDE effectively balances exploration and exploitation for complex optimization challenges.
• The algorithm shows significant potential for engineering design and computational intelligence applications.