An Improved Real-Coded Genetic Algorithm Using the Heuristical Normal Distribution and Direction-Based Crossover

Area of Science:

• Computational intelligence
• Optimization algorithms
• Engineering mathematics

Background:

• Constrained optimization problems are prevalent in engineering and science.
• Existing real-coded genetic algorithms face challenges with convergence speed and population diversity.
• Developing efficient algorithms is crucial for solving complex optimization tasks.

Purpose of the Study:

• To propose a Multi-Offspring Improved Real-Coded Genetic Algorithm (MOIRCGA) for constrained optimization.
• To enhance convergence speed and solution quality.
• To address limitations in existing genetic algorithm operators.

Main Methods:

• Development of a novel Heuristical Normal Distribution and Direction-Based Crossover (HNDDBX) operator.
• Introduction of a substitution operation to prevent population duplication.
• Implementation of a Combinational Mutation method for simultaneous local and global search.

Main Results:

• MOIRCGA demonstrated fast convergence speed across sixteen test examples.
• The algorithm effectively avoids population duplication, improving computational efficiency.
• Optimization of a cantilevered beam structure showed MOIRCGA outperformed the standard RCGA.

Conclusions:

• The proposed MOIRCGA, with its enhanced operators, offers significant improvements in solving constrained optimization problems.
• The algorithm achieves faster convergence and superior objective function values compared to existing methods.
• MOIRCGA shows promise for practical engineering optimization applications.