A Crisscross-Enhanced Groupers and Moray Eels Optimization Algorithm: Benchmark Test and Production Optimization | JoVE Visualize

Area of Science:

Computational intelligence
Optimization algorithms
Metaheuristic computing

Background:

Metaheuristic algorithms struggle with balancing exploration and exploitation, leading to suboptimal solutions on multimodal problems.
The Groupers and Moray Eels (GME) algorithm offers a cooperative framework but can lack topological coverage in constrained landscapes.
Existing methods face challenges in maintaining diversity and preventing stagnation in complex optimization tasks.

Purpose of the Study:

To enhance the Groupers and Moray Eels (GME) algorithm for improved performance in complex optimization scenarios.
To address limitations in global exploration, local exploitation, and topological coverage of existing metaheuristic approaches.
To introduce a novel algorithm, GPS-CC-GME, integrating deterministic initialization and dual-crossover search strategies.

Main Methods:

Developed GPS-CC-GME by incorporating a number-theoretic Good Point Set (GPS) for uniform initial agent distribution.
Implemented a dual-crossover search architecture: horizontal for global diversity and vertical for local fine-tuning.
Evaluated performance on the CEC2017 benchmark suite and a high-dimensional petroleum reservoir production optimization problem.

Main Results:

GPS-CC-GME achieved the highest overall ranking on the CEC2017 benchmark suite against baseline GME and other metaheuristics.
The enhanced algorithm demonstrated superior performance in a high-dimensional petroleum reservoir production scenario, yielding higher Net Present Value (NPV).
Results confirm the efficacy of deterministic initialization and bidirectional mutation operators in multipredator optimization models.

Conclusions:

GPS-CC-GME effectively overcomes the limitations of traditional metaheuristics in balancing exploration and exploitation.
The proposed enhancements significantly improve search efficiency and solution quality for non-linear engineering optimization tasks.
This work highlights the potential of integrating deterministic methods and advanced mutation strategies into cooperative search frameworks.