Supply chain integrated resource-constrained multi-project scheduling problem

Over the last few decades, supply chain integrated project scheduling problems have received greater attention to ensure profitability in organizations. Supply chain integrated project scheduling incorporates ordering project materials, conducting procurement from the suppliers, supplying materials to the warehouse, and executing project activities that create value in the integrated system. This paper presents a mathematical model and solution methods for a supply chain integrated resource constrained multi-project scheduling problem (SCIRCMPSP) with discounted cash flows. The mathematical model for the proposed SCIRCMPSP considers decisions regarding materials ordering (MO), supplier selection (SS), procurement, inventory, sharing of global resources among the projects, and completion of project activities within the deadline. A mixed integer programming (MIP) model is proposed for this SCIRCMPSP that aims to maximize the project's net present value (NPV). Five different meta-heuristic approaches are proposed to solve the model: immune algorithm (IA), genetic algorithm (GA), IA with forward-backward improvement (FBI) (IAFBI), GA with FBI (GAFBI), and surrogation-based GA (SGA). The performance of the meta-heuristics was tested on 72 self- generated small-to-large SCIRCMPSP instances ranging from 2 to 20 projects in a multi-project set, each with 30 to 120 activities. The experimental results demonstrate that the proposed SGA is significantly (p < 0.05) . 05 ) better than other algorithms in generating quality solutions. The effectiveness of the proposed SGA is further validated by solving classical multi-project benchmark instances. The results show that the proposed SGA is effective not only in solving SCIRCMPSPs but also in solving multi-project scheduling benchmark instances.