Content area
This work contributes to an important challenge faced by the semiconductor backend manufacturing industry with high operational complexity and energy consumption. Traditional NEH scheduling algorithms minimize makespan but do not consider the energy and sustainability objective. To address this, we introduced and evaluated a new hybrid algorithm, PIG_NEH (Population-based Iterated Greedy NEH), which incorporates iterative refinement and population based search methodologies. The method is based on computer simulations implemented in Python (v3. 11) on a Windows 11 machine powered by a 2.41GHz AMD Ryzen 3 CPU paired with 4GB of RAM. The experiments, through the use of static data, saw evaluations of scheduling performance for five distinct parameter configurations, including four (4) datasets with sizes of 8, 20, 30 and 50, using 5 and 20 machines. We evaluated metrics like makespan, energy efficiency and computational time. Statistical comparisons were made between NEH and PIG_NEH and shown as t-tests to highlight trade-offs. PIG_NEH reduces makespan by a further 1.85% over NEH as well as increases energy efficiency although a 37,622.54 % computational overhead is attained. These results contribute to sustainable scheduling practices in two ways, with the first being improved utilization of resources and the second being the integration of energy-saving goals into the manufacturing systems.