Intelligent sense response system for supply chain event management

More often than not, there are discrepancies between what was planned and what is/can be, or should eventually be executed. For instance, production shortages, missed deadlines, delays in receiving and dispatch of shipments, and order changes are more or less ubiquitous. If service level requirements (SLR) are associated with the successful implementation of the plans, these discrepancies between the planned and the executed can be expensive. In the business world, SLRs such as fill rates, on time delivery, zero defects, are associated with the successful implementation of the plans-and the discrepancies are expensive. The causes of these discrepancies can be attributed to poor planning or execution, or the occurrence of unplanned "events"---event management (EM) attempts to limit the losses incurred due to some of these "events" categories. The intelligent sense-response approach for EM focuses on the control theoretic approach.;Concepts developed within the realms of process control have proven that closed loop systems have superior performance in terms of maintaining the SLR and rejecting disturbances than the corresponding open loop systems. It is an extension of this concept, i.e., tighter integration between planning and execution through both feed forward and feedback mechanisms that the intelligent sense-response approach presented in this dissertation aims to exploit for performance enhancement of supply chains. Responding to the identified discrepancies is an important component of the proposed approach, and a bulk of this dissertation focuses on algorithms for efficient response. The rest of the dissertation focuses on the integration of the sensing capabilities that the supply chains are currently equipped with, with the proposed response mechanisms.;The first part of the dissertation proposes the aggregate high fidelity modeling (AHFM) approach for planning and replanning that provides evaluated responses to identified discrepancies with the objective of minimizing the losses as well as impact of the event on the SLRs of the system. AHFM aggregates the flexibility/redundancies that are available at various stages of planning and optimally assigns the resource/time from the aggregate pool to compensate the impact of the disturbance and problems. A solution approach for the AHFM approach, the computational complexity of which scales linearly or is polynomial is also presented. The second section of the dissertation discusses the integration of the AHFM approach with the sensing capabilities of the supply chain.;The formulation and solution of AHFM has been explored in several domains: manufacturing, air travel and food processing have been considered. Early investigation into these problem domains has yielded remarkable improvements in performance (up to 30% improvement in a discrete manufacturing environment, 84% improvement in process industries and up to 25% improvement in fleet scheduling). For the case study taken up for the IBM supply chain, the indicated improvements are of 27%. While these improvement percentages cannot be generalized for evaluating gains from implementing AHFM, the examples illustrate that at least in some cases, the magnitude of potential profits cannot be ignored---specially when this improvement comes at little extra cost. In addition, preliminary results of AHFM indicate that the technique yields feasible solution to problems that were considered infeasible.