The multiple resource constrained scheduling problem
Montes, Elliot J.
MetadataShow full item record
In this dissertation, algorithm methods that provide solutions to the Mutiple Resource Constrained Scheduling (MRCS) problem are presented. The MRCS problem can be considered an m machine, n job problem with batch setups subject to resource (worker) constraints with the objective of minimizing makespan, the maximum job completion time. Additionally, the machines are non-identical and parallel where job preemption is not allowed. The MRCS problem exists in systems where each job set requires processing by a specific machine and a crew of workers. Even though multiple machines exist in the system, not all machines can perform all jobs. Furthermore, before a worker(s) can be assigned to operate a machine for a particular job, the respective worker(s) must have been trained to perform that job. In addition, workers must be retrained before performing any new type of job. The goal is to schedule the jobs, workers, and machines such that the total makespan is minimized. This problem is unique because not only are workers and machines considered, but the workers have the additional job training requirements. The MRCS problem can be found in high health risk and environmental risk systems where continual operator training is mandatory before work can begin or continue. Such operations can be found in nuclear plant maintenance, defense industries, and other areas. The developed algorithm is evaluated via Monte Carlo sampling, existing decision rule comparison, and through human subjects testing.