A novel exact solution algorithm for a robust product portfolio problem under return uncertainty

Abstract. This research is aimed to address the optimization of a product portfolio problem under uncertainty using the principles of nancial portfolios theory. Since the success of a product portfolio is dependent on strategic decision making as well as on future changes of return, the return is best considered when it is deemed an uncertain parameter. The specic innovation of this research is the use of a robust optimization approach and providing an exact solution algorithm based on the model of Bertsimas and Sim. Given the uncertainty of the returns, the product portfolio model was developed based on the robust counterpart formulation of Bertsimas and Sim. An exact solution algorithm was also formulated to reduce the solution time. The results obtained by applying the model to a real case study of the dairy industry in Iran showed that increasing the condence level would decrease total returns of the portfolio and increase its total risk. A comparison between the proposed algorithm and similar methods showed that, on average, it would make 3% improvement in the solution time.