A Multi-Objective Coil Route Planning System for the Steelmaking Industry Based on Evolutionary Algorithms

In this paper a novel route planning system for steel coils that must pass through different processing steps of a generic steelmaking plant will be presented. Production times and costs are often the only considered indicators by traditional planning systems, while, with this newly proposed approach, customers’ quality requirements are also taken into account. In facts, in medium/large steelmaking plants there could be different processing line that perform the same processing step, but with different characteristics (e.g. in terms of flatness, crossbow, etc.). Therefore, the final quality of a coil greatly depends on the route it follows among the different processing lines. Moreover, over-quality, i.e. assigning a high quality coil to a less demanding customer order, must be avoided too. The proposed system is based on Multi-Objective Optimisation and, in particular, it can exploit different paradigms of Multi-Objective Evolutionary Algorithms. The planning system has been developed in C++ (for the optimisation module) and C# (for the graphical user interface). One of its key features is that it is highly configurable so that it can be easily adapted to several real industrial scenarios by means of simple XML configurations file describing the plant and the quality indicators to take into account during the optimisation process.