A dual-adaptive stochastic reinforcement chimp optimization algorithm for fire detection and multidimensional problem...

Area of Science:

• Computational Intelligence
• Swarm Intelligence
• Nature-Inspired Algorithms

Background:

• The Chimp Optimization Algorithm (CHOA) is a nature-inspired metaheuristic.
• Original CHOA faces challenges with slow convergence and local optima in multidimensional optimization.
• Addressing these limitations is crucial for practical applications.

Purpose of the Study:

• To propose a novel variant of CHOA, termed TASR-CHOA, to overcome existing limitations.
• To enhance convergence speed and improve exploration-exploitation balance in optimization.
• To validate the effectiveness of TASR-CHOA on diverse benchmark and real-world problems.

Main Methods:

• Developed TASR-CHOA by integrating a stochastic approach and a dual adaptive weighting mechanism.
• Evaluated TASR-CHOA on 29 conventional, 10 IEEE CEC-06, and 30 IEEE CEC-BC benchmark functions.
• Compared TASR-CHOA against 4 categorical and 18 IEEE CEC-BC algorithms using statistical tests.

Main Results:

• TASR-CHOA achieved superior performance, ranking first in 54 out of 73 evaluation functions and engineering problems.
• Demonstrated comparable results to state-of-the-art algorithms like SHADE and CMA-ES in several cases.
• Successfully applied TASR-CHOA to a computer-aided fire detection task using deep convolutional neural networks.

Conclusions:

• TASR-CHOA significantly improves upon the original CHOA for complex optimization tasks.
• The proposed enhancements lead to faster convergence and better global search capabilities.
• TASR-CHOA offers a robust and effective optimization tool for various scientific and engineering applications.