Optimization of metal utilization for aluminum extrusion operations

The production of aluminum billets by cutting long billets generates a lot of scrap, which can be very costly to recycle. When a cutting plan is applied to a billet, a last slug (scrap) often remains, the length of which does not correspond to any order. To minimize the amount of scrap, the problem is modeled and solved using mathematical programming.

In the aluminum industry, billets are among the most important products. These are long, solid cylinders of various alloy diameters, which have to be cut into billets of customer-specified lengths. These billets are then shaped by hot extrusion. The cutting process generates scrap that has to be remelted and recycled, generating additional production costs. Reducing scrap means improving productivity and cutting production costs.

In collaboration with a major aluminum producer, we analyzed the billet cutting process. This process is currently planned manually by a scheduler. It takes him several days to plan a billet cutting process that satisfies demand for only a few weeks. The resulting solution is far from optimal and generates a lot of scrap. What's more, if new orders come in after planning, it's very difficult for the planner to revise his initial plan.

The use of integer linear programming makes it possible to develop efficient mathematical models that lead to optimal or near-optimal solutions in a matter of hours or even minutes. This allows the planner to concentrate on more important tasks, and to gain the flexibility to integrate new orders after the plan has been built.