Sustainable retrofitting through multi-objective optimization: a time–cost–energy framework using opposition-based NSGA-III

Retrofitting projects are critical for enhancing the functionality, safety, and sustainability of existing infrastructures. However, selecting appropriate retrofitting strategies involves complex trade-offs between time, cost, and energy consumption. This study presents a comprehensive time–cost-energy consumption trade-off (TCECT) optimization model using the opposition-based non-dominated sorting genetic algorithm III (OBNSGA-III). The model considers multiple retrofitting aspects, each having several execution options defined by distinct time durations, costs, and energy usage. The objective is to simultaneously minimize project completion time, overall cost, and energy consumption while considering precedence relationships and limited discrete resources. OBNSGA-III integrates opposition-based learning (OBL) into the NSGA-III framework for improved initialization and generation jumping, enhancing convergence and diversity of solutions. A real-world case study with eleven retrofitting aspects demonstrates the model’s practical applicability. The results yielded 22 Pareto-optimal solutions, allowing decision-makers to select balanced trade-offs based on project priorities. Compared to other optimization models such as NSGA-III, MOPSO, and OB-MODE, the proposed approach shows superior performance across key indicators including hypervolume, generational distance, and solution spread. The findings affirm the model’s effectiveness in aiding sustainable, cost-efficient, and timely execution of retrofitting projects. This research contributes to the advancement of decision-support systems for infrastructure project planning and optimization.