Abstract
Scheduling is the process of organizing, choosing and timing resource usage in a facility (e.g. factory, hospital) to conclude all necessary activities for completing a job at a desired time. This is a common problem where there are multiple jobs and a variety of resources that are required to complete each one of them. The principal objectives in scheduling include efficient and effective resource utilization and improvement in overall system performance. Mathematical complexities in obtaining an optimal solution to scheduling problems force many assumptions into the existing solution. The importance and value of a customer or a job (based on historical data) are not included in such traditional solution procedures. In this paper, we present a decision support system (DSS) to solve scheduling problems that considers the value of a job and a customer. This real time scheduling system also permits many of the traditional assumptions and constraints of scheduling models to be relaxed. Schedule interruptions are permitted and the state of the facility is evaluated at the time of dispatching decisions. An expert system is used to generate alternatives in order to eliminate potential job lateness with reduced mathematical complexities of the scheduling algorithm. This DSS is developed as a tool for the decision maker to solve practical scheduling problems.
Traditional research on dynamic job shop scheduling (DJSS) is largely based on combinatorial analysis. Unfortunately, the NP-complete nature of the problem forces many assumptions that limit the use of analytical methods in practical problem solving to be included in existing solution procedures. Customer and job priorities, based on historical data, are seldom included in scheduling models. Many complexities in solving scheduling problems can be reduced if restrictions on limiting resources can be modified in consultation with the decision maker. Such changes can alter the feasible solution space, and permit sound, profitable decisions to enter into the model. This technique is not practicable in a traditional mathematical approach to obtain a solution for the scheduling problem. In this paper, we present a decision support system for real time scheduling that not only allows changes in limiting resources but also permits planned and unplanned downtime of machines, customer order changes and utilization of equipment with multiple capabilities. Expert systems and solicited user input are used to modify resource constraints. This system combines the customer value, job value and the potential value of job and customer along with a traditional component of scheduling analysis to create a schedule. The impact of this system on classical job shop scheduling is also discussed.