Approximate trade-off between minimisation of total weighted tardiness and minimisation of carbon dioxide (CO2) emissions in bi-criteria batch scheduling problem

The quantity of carbon dioxide (CO2) emissions is one of the most widely recognised measures of environmental sustainability. Given the mounting concern about climate change and global warming, managers are facing growing pressure to reduce CO2 emissions. In practice, other than CO2 emissions, managers may be concerned with other objectives when making a scheduling decision. This work develops the epsilon-archived genetic algorithm (epsilon-AGA) to examine two batch scheduling problems with the goal of minimising CO2 emissions and the traditional due date-based objective of minimising total weighted tardiness (TWT). Experimental results show that in terms of both quality and diversity of solutions, epsilon-AGA outperforms NSGA-II for same computation time limit as the stopping criteria. Several interesting observations are made. (1) These two objectives conflict with each other; (2) jobs that arrive soon after each other reduce makespan, and so reduce CO2 emissions; (3) given a set of m identical batching machines, the due dates of jobs do not seem to substantially influence CO2 emissions; and (4) in purchasing a machine, the variation in power consumption among machines is critical to reducing the TWT.