How best to schedule production work against the likelihood of a machine breakdown is being studied at The Hong Kong Polytechnic University.
An example of so-called preemptive-repeat work loss is in metal refining, said Principal Investigator Dr Zhou Xian. “The raw material must reach a specified temperature to obtain desired purity. If the process is interrupted by a power outage, the material cools rapidly, work done prior to the breakdown is lost and the process must start again.”
Dr Zhou and his researchers are tackling the problem theoretically as well as numerically, to determine optimum scheduling of different jobs under various criteria against the uncertainties of random breakdowns.
“We know it makes sense to schedule the shortest jobs first,” he said. “This way we can reduce the waiting time for subsequent jobs and therefore reduce proportional costs for items like inventory management.”
Other costs and factors such as penalties due to late delivery of work are also built into the formulas.
Said Dr Zhou: “We haven’t limited the research to a particular process but we’re coming up with analytical solutions for certain situations and algorithms to work out optimal schedules.”
Preemptive-repeat breakdowns are more difficult to analyse than preemptive-resume models where work is not lost because more factors are involved, said Dr Zhou.

Principal Investigator
Dr Zhou Xian : maxzhou@polyu.edu.hk