The quest for innovative solutions in optimization has led to an exciting development where Large Language Models (LLMs) are utilized to design novel metaheuristic algorithms. Riding on the capabilities of ChatGPT-3.5, researchers introduced the zoological search optimization (ZSO) algorithm, which is inspired by the collective behavior of animals in nature.

Highlights:

• Metaheuristic Algorithm Design: ChatGPT-3.5 assists in creating ZSO, which includes prey-predator interaction and social flocking operators.
• CRISPE Framework: The standard prompt engineering framework CRISPE influences the prompt design, guiding the development of the algorithm.
• Performance Analysis: ZSO and its variants were rigorously tested against CEC2014 and CEC2022 benchmark functions, as well as six engineering optimization problems.
• Competitive Edge: ZSO algorithms demonstrated efficiency and efficacy, facing off against 20 popular MAs with successful outcomes.

Incorporating AI into algorithm design signifies an unprecedented direction for the metaheuristics community. This paper lays the groundwork for future research, suggesting that LLMs like ChatGPT-3.5 can not only contribute but potentially revolutionize the metaheuristic problem-solving landscape.