Vittal Prabhu
Industrial & Manufacturing Engineering, Pennsylvania State University
This presentation will provide an overview of a methodology for controlling discrete-event systems by formulating these problems as continuous variable feedback control problems resulting in a unified mathematical and computational framework. The science-base of this work includes theories of discontinuous differential equations, Lyapunov stability, adaptive control, and nonlinear control theory, which can be used to prove convergence properties and to characterize emergent behavior of the resulting control systems. This has resulted in scalable parallel/distributed algorithms for a variety of applications including Just-in-Time production scheduling, maintenance scheduling, batch sequencing, inventory control, transportation, and distributed supply chains. Computational complexity of these control theoretic algorithms typically increases linearly with the number of events to be controlled, and worst case is of the order O(n²).

We will discuss some of the salient features of this approach along with applications. In the manufacturing context, we will discuss how the approach can be used for simultaneously controlling production, capacity, maintenance, and energy. In the context of services, we will discuss how the approach can be used for just-in-time delivery for green fleets as well as crowdsourced delivery services. In the later part of this presentation we will present some of our recent and ongoing efforts addressing ways to smarten service systems using techniques from operations research in our SEE 360 initiative at Penn State. The presentation will conclude with some thoughts on integrating OR and AI techniques with this feedback control problem as well integrating IIoT sensors for improving fidelity.