| Popis: |
Processes and markets are subject to uncertainty which makes production activities of batch plants constitute a complex environment to manage. Uncertainty may cause deviations and infeasibilities in projected schedules; this can lead to poor planning and inefficient use of materials. Consequently, the relevance of explicitly incorporating variability in the scheduling formulation in order to provide more efficient plans and robust decisions to changes has been recognized. Moreover, a significant performance improvement of the SC can be obtained by comprising low level decisions in managing the supply chain. Also, such an integration results in a better treatment of supply chain dynamics. However, the inclusion of scheduling models lead to large scale problems. Therefore, a challenge in this field is the reduction of the computational burden required to solve this kind of problems. This chapter addresses the batch plants scheduling under exogenous uncertainty. The most widely utilized approach to tackle this problem is stochastic programming; however, its solution is at the expense of a significant increase of computational cost. From another standpoint the S-graph representation, a graph-theoretic approach, has proved to be very efficient to deal with deterministic scheduling. In this chapter, the S-graph framework is enhanced so that stochastic scheduling problems can be handled. For this purpose, a Linear Program (LP), which serves as performance evaluator, has been coupled with the S-graph framework. One of the main advantages of the proposed approach is that the search space does not exponentially increase according to the number of scenarios considered in the problem. The potential of the proposed framework is highlighted through two illustrative examples. |
