A multiperiod model for a training program

This paper is concerned with formulating and optimizing a multiperiod planning model for a manpower training program. Firstly, a time-dependent feed-forward model which traces eligible individuals through an educational (training) program and ultimately to a state of employment or unemployment is developed. The objective function is composed of an educational cost component and a manpower shortage or surplus cost component. Thus, the student (or trainee) optimal acceptance policy is derived. It is also proven that a one-state variable dynamic programming algorithm is much more efficient than a two-state variable algorithm in arriving at the optimal solution.