Task time estimation in a multi-product manually operated workstation

Many manual workstations are designed so that a specific set of tasks to be completed on jobs are performed at the workstation. The subset of individual tasks executed on each job may vary, and a jobs total processing time at the workstation is the sum of those individual task times. Since estimates of the mean and variance of the individual task times are used to make operational and planning decisions, data should be collected on a regular basis to ensure accurate estimates. In this research, we apply a least-squares method and maximum likelihood estimation to estimate the mean and variance of individual task times at a manual workstation from total job-processing time data. Both methods assume that the time to execute individual tasks at a workstation can be treated as independent random variables whose distributions are the same regardless of what other tasks are executed on a job. The maximum likelihood method also assumes that these times are normally distributed. Efficient computational formulas developed for the least-squares method are ideal for use with an automatic data collection system, and both methods provide good estimates of mean task times that are adequate for planning and operational decisions.