|M.Sc Student||Nir Michael|
|Subject||A Profit Based Algorithm for Managing a Dynamic, Resource|
Constrained Multi-Project Environment
|Department||Department of Industrial Engineering and Management||Supervisor||Professor Avraham Shtub|
The research examines a heuristic establishing a project acceptance policy in a dynamic multi-project resource constrained environment. Past research has focused on time related project acceptance policies limiting the number of projects to a predefined level. In this paper, we present a Dynamic Programming formulation for selecting profitable projects. Project profit is defined as the capitalized income and expense from an accepted project. Due to the complexity of the problem, we propose a heuristic based on time in system estimation. The estimation is realized through implementation of the snap shot principle adapted from queuing theory.
We propose hypotheses for evaluating our approach in a wide range of treatments. We examine the proposed method efficiency and effectiveness in three independent experiments. Full factorial designs are used .Our method maximizes the profitability compared with time based project acceptance and random project acceptance policies. Through implementation of a backlog list, control and management of project load and sensitivity analysis of resource load can be accomplished. We show that project acceptance rules dictate performance. In addition, by selecting profitable projects our method establishes an adaptive window approach; the number of projects in the system fluctuates according to the system state. In dynamic stochastic setting an adaptive window is superior to a fixed window approach.
The research offers a management tool for a multi-project dynamic environment. Validation of the results obtained in the study through implementation in an industrial project management environment, can open the door for future application of our method.