Resource-constrained project scheduling problem with hybrid energy and dynamic energy prices

This paper addresses renewable energy, the fluctuating characteristics of time-of-use (TOU) tariffs and multi-shift scheduling by proposing a novel model for the resource constrained project scheduling problem with multi-shifts, dual-energy hybrid and TOU (RCPSP-MS-DH-TOU) to minimize total project costs. To achieve this, the study focuses on the coordinated optimization of different energy resources during project scheduling and the strategic allocation of high-energy-consuming tasks to pe riods of lower electricity prices. A mixed-integer linear programming model is devel oped to describe the RCPSP-MS-DH-TOU, with an algorithm called MD3QN-PER introduced to solve it. The MD3QN-PER algorithm combines an exact algorithm with an enhanced multi-step prioritized experience replay dueling double deep Q network, which learns optimal scheduling strategies by training on mixed instances. Additionally, spatial pyramid pooling enhances the model’s generalization capabil ity, enabling it to handle scheduling problems of different scales. The experimental results demonstrate that the proposed algorithm outperforms widely used methods in both effectiveness and efficiency. Specifically, for large-scale problems, the algo rithm effectively balances the quality of the solution and the computation time. The application of the RCPSP-MS-DH-TOU model enables project managers to develop more efficient scheduling plans by considering renewable energy, TOU tariffs and multi-shift scenarios.