In this paper, we study multi period project portfolio selection problem. To satisfy more realistic assumption and solve real world cases we develop a new mathematical model which considers risk issues, stochastic incomes and possibility of investigating extra budget of each period out of portfolio such as investigating in bank. The main objective of this model is to maximize net earned profits. Due to complexity of problem, exact methods could not solve it in logical times, so a new genetic based algorithm with new solution representation and operators is developed; it is then compared with a well-known algorithm so called PSO. The performance of these algorithms compared by the means of two prominent indicators called hypervolume and spacing metric. Computational confirm the efficiency of the proposed method.
- Portfolio selection,
- risk analysis,
- genetic algorithms,
- project interdependency
Available at: http://works.bepress.com/ali_asghar_tofighian/3/