The way orders are processed on any production line can have a significant impact on cost, productivity and quality, and the reliability of the facility itself. This is especially true for steel production lines, and a challenge that dates back many, many years.
Until recently, production scheduling challenges were addressed, to some extent, by leveraging operator experience or using constraint- and rule-based commercial systems. However, that method was insufficient.
First, the dimensionality of the problem must be considered. There are multiple relevant properties of a product, and they are confronted many times in terms of cost to reach the global optimal schedule. All of them must follow adequate trajectories along the sequence at the same time. Typically, one change made to improve one issue would have an adverse effect somewhere else.
Second, the size of the search space grows with the factorial of the number of items. Did you know that the number of potential sequences of just 70 items, or a typical day's worth of production in a galvanizing line, is more than 10^109. To put this into perspective, consider that the lifetime of Earth measured in seconds is in the order of 10^17. The number of atoms in the observable universe is estimated at around 10^80. In order to explore all possibilities in an exhaustive way, the best computer in the world would need many years to solve these problems using traditional approaches. It’s no wonder the traditional methods of addressing production scheduling found only limited success!
ArcelorMittal Global R&D developed a novel approach to analyzing and addressing scheduling challenges by relying on mathematical modelling of the effects of the ordering and 'bio-inspired algorithms' that intelligently explore the search space. Bio-inspired computing is an emerging approach that represents the advances in computer science, mathematics and biology over the years. It relies on the principles and inspiration of the biological evolution of nature to solve computer science problems.