Novel Mixed-Integer Linear Programming Model for the Multiproduct Pipeline Network Scheduling Problem

The pipeline is one of the most reliable and economical modes of transport. It is mainly used for petroleum products transportation over long distances. In this paper, we are interested in tackling the multiproduct pipeline network scheduling problem in the context of the mining industry. The pipeline network under study is part of the OCP Group’s pipeline network for the transportation of phosphate ore products. It is composed of a main pipeline and several secondary pipelines. We aim to schedule a set of batches of products to be transported from the main pipeline to intermediate storage groups at the terminal station before being delivered to clients using secondary pipelines. Each batch can be composed of several blocks, and each block is destined to one storage group at the terminal station. A novel Mixed-Integer Linear Programming (MILP) model based on continuous time and batch volume formulation is proposed. The objective is to maximize the main pipeline utilization rate for phosphate ore transportation with the smallest number of blocks to satisfy all clients’ demands. A set of instances is generated from the pipeline network of OCP Group and numerical experiments prove that the proposed model gives effectively optimal solutions for instances of short-term planning horizon, and near-optimal solutions for instances of long-term planning horizon.