This paper models a realistic problem involving workforce assignment and training for a large manufacturing environment. In this particular environment, the workforce is undertrained and most assignments will result in necessary training. This problem was previously addressed as a single objective problem. This paper expands to a multi-objective formulation. This is a more accurate reflection of the problem because almost all real world problems have many objectives which can be conflicting.

The program developed in this paper is designed for use by supervisors in the production setting. A two stage program is designed where the first stage generates initial solutions by solving each objective function idependently of the others. Meta-RaPS—a modified greedy algorithm—is used to find these solutions. The user selects one of these solutions to carry into the second stage: compromise programming. The second stage uses input from the user in an iterative and intuitive fashion. This input guides the program to the solution which the user determines is the best compromise solution.

Meta-RaPS is effective at finding a good solution extremely quickly. There is an important trade-off between the quality of solutions and computational run-time which will need to be tweaked for a specific application. The compromise programming stage could benefit from coding improvements; however, it is still effective at allowing the user to guide the program towards the best compromise solution by assigning trade-off values between objectives.