A Selective Biogeography-Based Optimizer Considering Resource Allocation for Large-Scale Global Optimization.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• Biogeography-based optimization (BBO) is efficient but struggles with high-dimensional problems.
• The curse of dimensionality significantly degrades BBO's performance as search space increases.

Purpose of the Study:

• To improve Biogeography-based optimization (BBO) for large-scale global optimization problems (LSOPs).
• To introduce a selective migration operator (SBBO) addressing BBO's dimensionality sensitivity.
• To enhance cooperative coevolution (CC) with efficient resource allocation for LSOPs.

Main Methods:

• Developed Selective Biogeography-based Optimization (SBBO) with heuristic selection of differential and normal distributed migration operators.
• Integrated cooperative coevolution (CC) strategy to tackle large-scale optimization challenges.
• Proposed an efficient computing resource allocation method to address subgroup imbalance in CC.

Main Results:

• SBBO demonstrated superior effectiveness and efficiency compared to BBO variants and other algorithms on the CEC 2010 LSOP benchmark suite.
• Extensive experiments validated the performance gains of SBBO for large-scale global optimization.
• The proposed resource allocation strategy proved vital for optimizing within limited computational budgets.

Conclusions:

• SBBO effectively overcomes the curse of dimensionality inherent in BBO for LSOPs.
• The combination of SBBO and cooperative coevolution offers a powerful approach for large-scale optimization.
• Efficient resource allocation is crucial for successful large-scale optimization under computational constraints.