A Comparative Study of Multi-Objective Optimization Algorithms for Chiller Selection in Cooling Plants

The optimal selection of chillers in cooling plants for buildings and industrial facilities remains a critical challenge in the heating, ventilation and air-conditioning (HVAC) industry. This study aims to comparatively evaluate different algorithms for solving the chiller selection problem as a multi-objective problem considering capital cost and energy consumption. A case study is conducted using a cooling plant with a known load profile and a set of 13 air-cooled screw chiller models. Four algorithms (NSGA-II, MOEA/D, IBEA, and PBEA) are applied and assessed based on their ability to approximate the Pareto front. The results indicate that NSGA-II outperforms the other algorithms, contributing 81.8% of the reference Pareto front solutions and achieving the best performance metrics. Additionally, a representative trade-off solution demonstrates a 16.4% reduction in energy consumption with only a 2.0\% increase in capital cost, highlighting the effectiveness of multi-objective optimization in supporting efficient HVAC system design decisions.