Abstract
This paper deals with a problem of optimization of resource allocation in a project plan. It is assumed that the technological ordering of activities in a project is defined by the project graph. The predecessor-successor relations of activities are mapped into a set of transformations, where the state variables are the amounts of work required to complete the activities. The resources allocated to each activity are considered as decision variables. They are constrained when resources are limited. This enables one to consider the project execution as a discrete multistage decision process with the associated objective function. The optimization of this process is viewed as a dynamic programming problem. It is shown that the high dimensionality of the functional equations that appear can be overcome by means of (a) subproject programming, (b) two-level concept of management control evaluation, and (c) by approximation in policy space.

This publication has 0 references indexed in Scilit: