Optimization of time–cost–quality-CO2 emission trade-off problems via super oppositional TLBO algorithm

Abstract

The teaching–learning-based optimization (TLBO) algorithm is widely recognized for its efficiency and effectiveness in solving optimization problems. However, it often encounters challenges with premature convergence, leading to local optimal solutions. To address this limitation, this study introduces an enhanced variant of TLBO, denoted as super oppositional teaching–learning-based optimization (SOTLBO) algorithm. This enhancement introduces a novel super opposition learning (SOL) strategy, which retains superior candidate solutions by simultaneously evaluating an individual and its corresponding opposite individual. The proposed SOTLBO is applied to a time–cost–quality-CO₂ emission (TCQCE) trade-off problem involving a 33 activity project that considers all logical dependencies among activities. Results demonstrate that SOTLBO achieves faster convergence and higher-quality optimal solutions. To assess the algorithm’s effectiveness, its performance is compared with well-established algorithms: slime mold algorithm opposition tournament mutation (SMOATM), golden ratio sampling based random oppositional aquila optimization (GRS-ROAO), and plain TLBO algroithms. Statistical analysis highlights that SOTLBO outperforms these algorithms, achieving the highest hyper-volume (HV) value of 0.889 and the suitable mean ideal distance (MID) and spread (SP) values of 1.918 and 0.382, respectively, for the 33 activity project. These findings highlight SOTLBO’s superior ability to enhance diversity and ensure more uniform solution distributions compared to other multi-objective evolutionary algorithms.