An Adaptive Smell Agent Optimization with Binomial Crossover and Linnik Flight for Engineering Optimization Problems

Optimization problems are prevalent in engineering, often requiring effective methods to navigate complex, high-dimensional landscapes with multiple local minima. Existing algorithms frequently fall short due to limitations in handling diverse constraints and complexities. This paper proposes the adaptive crossover-based smell agent optimization (ACB-SAO) algorithm inspired by the olfactory sense in living organisms. The new algorithm introduces two key contributions, i.e., a longtail exploring mode integrating Linnik Flight with a golden ratio configuration to improve exploration capabilities and a dynamic crossover rate adjustment for smell agent optimization (SAO). This synergy enhances solution accuracy by balancing global and local search capabilities. To validate its performance on complex numerical benchmarks and engineering design problems, ACBSAO is compared with seven well-known and recent competitive algorithms on 23 classical, 29 CEC2017, 30 CEC2022 benchmark functions, and 14 real-world engineering design problems. The results in a scoring system indicate that ACB-SAO achieved the maximum score of 100 for the CEC2017, CEC2022, and realworld engineering designs, demonstrating that it outperforms other algorithms and significantly improves upon the standard SAO. These results highlight ACB-SAO’s potential in solving practical optimization problems, proving its effectiveness and advantages in addressing complex challenges.