Load optimization of cogeneration units based on intuitive multi-objective fish swarm algorithm

• In this paper, a multi-objective optimization algorithm is proposed to solve the optimization of complex problems such as cogeneration loads. • The introduction of intuitionistic fuzzy entropy to the multi-objective fish swarm algorithm enhances the monitoring of the algorithm’s operational status and the algorithm’s ability to find optimal accuracy. • The multi-objective fish swarm algorithm has demonstrated its adaptability and effectiveness in optimizing solutions to certain engineering problems. This study addresses the multi-objective optimization challenges in seasonal heat-power load distribution for cogeneration units by proposing a multi-objective artificial fish swarm algorithm based on intuitionistic fuzzy entropy (IFEMOAFSA). The algorithm enhances the original intuitionistic fuzzy entropy framework, integrating membership, non-membership, and hesitation degrees to guide fish swarm behavior. It dynamically categorizes swarm particles into three states, improving solution space coverage and priority-based solution identification. Convergence direction is adaptively adjusted using intuitionistic fuzzy entropy, with Pareto frontier solutions determining optimal load allocation. Evaluated via the Zitzler-Deb-Thiele (ZDT) benchmark functions, IFEMOAFSA achieves a 42.63% comprehensive performance improvement over four benchmark algorithms, verified by Mean Inverted Generational Distance (MIGD) and Mean Hypervolume Metric (MHV). A cogeneration unit model incorporating operational characteristics and historical data demonstrates the method’s efficacy: multi-objective balance is maintained across iterations, achieving a 1.41% thermoelectric load increase and 1.54% optimal coal consumption reduction. The algorithm reduces heat/electricity losses and operational costs under diverse conditions while enhancing load utilization rates. These results validate IFEMOAFSA’s effectiveness in solving annual load optimization challenges for cogeneration systems, showing promising applications for similar multi-objective optimization problems requiring dynamic adaptability and robust convergence properties.