An Expert System for Predictive and Reactive Scheduling of Multiproduct Batch Plants

HENNING, GABRIELA; CERDÁ, JAIME

LATIN AMERICAN APPLIED RESEARCH

PLAPIQUI(UNS-CONICET)

Lugar: Bahia Blanca; Año: 1995 vol. 25 p. 187 - 198

0327-0793

A knowledge-ased system has been developed to address the short-term scheduling problem of an important segment of multiproduct bath plants featuring a single bottleneck processing stage with a set of parallel units operating in a campaign basis mode. An Adequate operation of such plants involves the fulfillment of an important number of customer orders for different products to be manufactured in the same facility using a set of limited resources such as processing equipment, raw-materials, utilities and manpower. To aid industry to cope with the problem, the system provides a decision support environment for predictive and reactive production planning in which the scheduler can efficiently generate, explore and appropriately compare alternative solutions. Graphical representations such as Gantt charts and stock evolution line-plots reporting tardy campaigns and raw-material shortages help assess the goodness of a given schedule. To make it highly modular and easy modifiable, the system has been implemented in the KAPPA-PC (Intellicorp, 1993) development environment using an object/oriented representation and a task-oriented problem-solving methodology.