Abstract:
Scheduling is a key factor for delivering a quality and reliable product. The sudden interest
in scheduling problems over the past forty years emphasizes the new opportunity created by
using scheduling tools (K r and Yazgan 2016). Scheduling tools enable companies around the
world to minimize non-value added factors like setup times, setup cost and changeovers. On
time delivery of products are achieved by optimizing the scheduling of production,(Gupta
and Chantaravarapan 2008). This project focuses on a wheat mill in Silverton, Gauteng.
This report considers a parallel machine scheduling problem, with sequence dependent setup
times for the production of
our products. The total demand of each job must be processed
at the same time, not allowing preemption. The primary objective of the schedule is to
minimize the total production time.
A Mathematical programming formulation shall form the basis of solving the problem.
Five heuristic rules are used. Results were obtained by running all of the heuristic rules over
thirty random demand scenarios. The optimal heuristic rule was determined as the process
with the most robustness to change in input data. In this project the largest
ushing times
heuristic rule performed the best.
The heuristic chosen as the best can easily be implemented by the company. No additional
resources have to be bought. The solution have been tested against real world data and
delivered excellent results. The current run time for the best heuristic rule is 0.0005 seconds.
The current scheduling method will schedule all the demand in approximately 23.9 days.
The new heuristic rule scheduling method will be able to produce all the demand in just 18.73
days. The nancial impact of implementing the optimal heuristic rule saves the company
up to R653.00 on electricity, R430.00 on water and R18 000.00 on overtime per day. That
lead to a total savings of R19 083.00 per day. The new heuristic will eliminate four days of
production. Equaling the total savings to R76 332.00 for four days.
