Tackling tough industrial challenges
Earlier this month I was at Carnegie Mellon University attending the Center for Advanced Process Decision-making (CAPD) Annual Review Meeting, at the kind invitation of Professor Ignacio Grossman. CAPD brings together industry practitioners (primarily from process industry organizations) and academia, and harnesses a lot of brain power to work on some tough industrial optimization and process systems challenges.
As well as formal presentations from CAPD members, there were poster sessions where graduate students working at CAPD described their work and answered questions. I was very impressed by the broad range of issues being addressed, by the enthusiasm of the students, and by the way that they all were able to explain their research: although there were enough mathematics and algorithm overviews displayed to make my brain hurt, each student was able to provide a close-to-layman’s terms explanation of their work (at least, close enough for me). You can view some of the presentations (alas, without the eloquent explanations) on the CAPD web site.
What struck me, as I listened to the presentations on optimization, and talked to the students later on, was that the focus of attention was all on optimization alone. I asked about disruption management several times, but nobody had a really satisfactory answer. In a “slow loop” planning cycle, where there’s time to work out a highly-optimized solution, and to absorb changes and disruptions that may crop up, it’s reasonable that if the operational situation changes then you simply rerun the optimization process and build a new solution from scratch, as many times as the planning cycle allows. But in the “fast loop” operational scheduling world, where the situation can change by the hour (and sometimes more frequently), it’s not feasible to keep rebuilding your solutions from the beginning, since this can have an adverse impact on activities already underway.
Actenum’s technology has been designed to provide both a high degree of optimization and disruption management capabilities (when a situation changes our solutions focus the rebuilding effort on the portion of the plan or schedule that are affected, leaving everything else alone as far as possible). Why, then, was there this gap in capability in the CAPD work? The answer, as far as I can tell, lies in the degree of problem complexity and the time available to the students: they can’t work on everything and so the first (and only) thing they focus on is the optimization aspect.
CAPD is an interesting vehicle for working on complex industrial situations, and warrants a close look from any process organization facing difficult challenges in their production environment.
