Environment Sensitivity-Based Cooperative Co-Evolutionary Algorithms for Dynamic Multi-Objective Optimization.

Area of Science:

• Optimization
• Computational Intelligence
• Evolutionary Computation

Background:

• Dynamic multi-objective optimization problems (DMOPs) present significant challenges due to conflicting objectives that change over time.
• Existing algorithms struggle to efficiently adapt to the temporal dynamics inherent in DMOPs.

Purpose of the Study:

• To propose a novel cooperative co-evolutionary strategy tailored for solving DMOPs.
• To enhance the adaptability and convergence speed of optimization algorithms in dynamic environments.

Main Methods:

• A new method for grouping decision variables based on environmental interrelation.
• Utilizing two populations for cooperative optimization of variable subcomponents.
• Employing differential prediction and Cauchy mutation for rapid environmental response.
• Integrating the strategy into NSGA-II and multi-objective particle swarm optimization to create DNSGAII-CO and DMOPSO-CO.

Main Results:

• The proposed algorithms, DNSGAII-CO and DMOPSO-CO, demonstrated superior performance compared to three state-of-the-art algorithms on seven benchmark DMOPs.
• Significant improvements in convergence and distribution were observed across most tested DMOPs.
• The cooperative co-evolutionary strategy effectively addresses the challenges posed by time-varying objectives.

Conclusions:

• The developed cooperative co-evolutionary strategy based on environment sensitivities is effective for solving DMOPs.
• The proposed DNSGAII-CO and DMOPSO-CO algorithms offer significant advantages in handling dynamic and multi-objective optimization challenges.
• This research contributes a robust framework for tackling complex, time-varying optimization problems.