Much of the literature on cross-training and worker assignment problems focus on simulating production systems under cross-training methods. Many have found that for specific systems some methods of allocating workers are better performing than others in terms of overall productivity and ability to deal with change. This has lead researchers to create mathematical programming models with a goal of finding optimal levels of cross-training by changing worker allocations. Learning and forgetting curves have been a key method to improve the solutions produced by the optimization models, but learning curves are often nonlinear causing increased solving times. Because of this, most works have been restricted to modeling small, simple production systems.
This thesis studies the expansion of worker allocation models with human learning and forgetting to include variable work structures, thus allowing the models to be used to address a larger set of problems than previously possible. A worker assignment model with flexible inventory constraints capable of representing different production structures is constructed to demonstrate the expansion. Utilizing a reformulation technique to counteract the increased solve times of learning curve incorporation, the scale of the production systems modeled in this work is larger than in similar works and closer to the scale of systems seen in industry. Production systems with multiple products and corresponding due dates are modeled to better represent the production environment in industry. Investigative tests including a 2^4 factorial experiment are included to understand the performance of the model. The output of the optimization model is a schedule of worker assignments for the planning horizon over all of the tasks in the modeled system. Production managers could apply the schedule to their existing lines or run what-if scenarios on line structure to better understand how alternative structures may affect worker training and line productivity over the planning horizon.
Library of Congress Subject Headings
Manpower planning--Data processing; Manufacturing industries--Management--Mathematical models; Employees--Training of--Mathematical models; Industrial productivity
Department, Program, or Center
Industrial and Systems Engineering (KGCOE)
Scott E. Grasman
Michael E. Kuhl
Chacosky, Austin T., "Allocation of Workers Utilizing Models with Learning, Forgetting, and Various Work Structures" (2015). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus