Nonuniform time discretization approach to batch and continuous process scheduling
Abstract
The scheduling of batch and continuous operations has received considerable attention in the recent process systems engineering literature. Two basic conceptual models have been investigated: the cyclic, campaigned operations type and the non-cyclic type. The former has been used for multipurpose operations in which cross-contamination is a major concern while the latter is found in many types of specialty chemicals settings. From a modeling point of view, the campaigned operations offer a more convenient representation of time via campaign lengths and cycle times, while the noncyclic operation requires that time be treated as a continuum. The by now classical approach to resource constrained scheduling problems is to introduce a discretization of time and a set of assignment variables which identify whether or not a particular task is to be executed in a particular equipment item in given time interval. In general to adequately represent the processing times and other event times of the plant, the time quantum required for the discretization can be quite small compared to the planning horizon length, thus leading to very large 0-1 variable dimensionality and often excessively large computation times. In this thesis a general framework for accommodating a wide range of scheduling scenarios arising in multi product/multipurpose batch/continuous chemical processes is developed. The scheduling problem is formulated as a mixed integer nonlinear program based on a continuous time representation. Flexible equipment assignment and variable batch sizes are taken into account. To solve the resulting MINLP model an algorithm based on the Bayesian Heuristic Approach (BHA) to discrete optimization was developed. Test results suggest that the BHA combined with the continuous time representation shows promise for the solution of batch/continuous scheduling problems.
Degree
Ph.D.
Advisors
Reklaitis, Purdue University.
Subject Area
Chemical engineering|Industrial engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.