Parameter setting of meta-heuristic algorithms: a new hybrid method based on DEA and RSM

Area of Science:

• Optimization techniques
• Computational intelligence
• Operations research

Background:

• Parameter tuning is crucial for meta-heuristic algorithm performance but is often time-consuming and experimental.
• Existing methods for parameter optimization can be inefficient, impacting overall algorithm effectiveness.

Purpose of the Study:

• To develop a novel hybrid method for optimal parameter selection in meta-heuristic algorithms.
• To simultaneously optimize parameters and maximize efficiency using the proposed method.
• To evaluate the effectiveness of the new method on standard and real-world problems.

Main Methods:

• A hybrid approach combining Data Envelopment Analysis (DEA) and Response Surface Methodology (RSM), termed DSM.
• Application of the DSM method for parameter setting of the Cuckoo Optimization Algorithm (COA).
• Testing on standard Ackley and Rastrigin functions and a real-world reverse logistics problem for COVID-19 waste management.

Main Results:

• The DSM method demonstrated superior performance compared to other approaches.
• Improvements were observed in solution time, number of iterations, efficiency, and objective function accuracy.
• The method proved effective for both standard test functions and complex real-world applications.

Conclusions:

• The hybrid DSM method offers an efficient and effective approach for optimizing meta-heuristic algorithm parameters.
• This method enhances both parameter optimization and overall system efficiency.
• DSM shows significant potential for application in diverse optimization challenges, including environmental management.