A novel metaheuristic optimizer GPSed via artificial intelligence for reliable economic dispatch

Area of Science:

• Engineering and Applied Sciences
• Computational Intelligence
• Optimization Techniques

Background:

• Meta-heuristic optimization algorithms are widely used but suffer from local optima, slow convergence, and high computational demands.
• Existing optimizers often struggle with complex, multi-variable problems common in engineering and scientific research.

Purpose of the Study:

• To introduce a novel, simple, and effective hybrid optimizer to address the deficiencies of current meta-heuristic algorithms.
• To validate the proposed hybrid approach across a range of meta-heuristic optimizers, from established to recent.
• To demonstrate the solver's applicability to real-world optimization challenges, specifically Economic Dispatch.

Main Methods:

• Developed a hybrid optimizer integrating an ambiguous optimizer with Artificial Intelligence (AI).
• Evaluated the hybrid optimizer's performance using Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Teaching-Learning-Based Optimization (TLBO), and Artificial Gorilla Troops Optimization (AGTO).
• Applied the hybrid solver to the Economic Dispatch problem of the IEEE 30-bus system.

Main Results:

• The proposed hybrid optimizer consistently converged to the global optimum solution.
• Achieved the minimum energy cost for the Economic Dispatch problem.
• Demonstrated superior reliability and adequacy with minimal iterations and computational requirements compared to individual optimizers.

Conclusions:

• The novel hybrid optimizer effectively overcomes the limitations of traditional meta-heuristic algorithms.
• The technique proves reliable and efficient, offering a robust solution for complex optimization tasks.
• Validated applicability across diverse meta-heuristic algorithms and practical engineering problems.