ABSTRACT
Purpose and Objectives
The purpose of this research is to present an integer linear programming (ILP) model that can
optimize the railway blocking. It determines the best assignment of shipments to blocks and
routes them through classification yards, replacing traditional decision-making with a
data-driven approach.
Methodology
The model uses clearly defined decision variables and constraints to control shipment assignment
and block creation. Implemented in AMPL (A Mathematical Programming Language) and
solved with Gurobi Optimizer, it balances two distinct costs in railway blocking, which are
distance traveled and handling (reclassification) costs while keeping blocking capacity and
shipment capacity limitation.
Scope of the Work
This study focuses mainly on the blocking aspect of railway operation and specifically, the
assignment of shipments to blocks and their routing. It does not address block-to-train
assignments[1] or train scheduling, allowing a more focused examination of cost reduction in
blocking.
Results
Testing on two benchmark networks, a simple five-yard network, and a larger-more complex
regional network with 39 nodes. From the experimental result, it showed that the ILP model
outperformed traditional expert plans. In the larger network, the optimal solution provides an
85% reduction in total costs, however, to produce the most optimized blocking plan, the ILP
requires nearly 12 days to complete.
Conclusions and Recommendations
The results from the two network experiments show that the ILP model significantly improves
the cost-effectiveness of blocking plans, with even sub-optimal solutions can provide a
noticeable savings over traditional methods. However, a major drawback is its long
computational time for large-scale networks. Obtaining an optimal solution can be impractical
for real-time decision-making[2].