This paper is a novel work of portfolio-selection problem solving using multi
objective model considering four parameters, Expected return, downside beta
coefficient, semivariance and conditional value at risk at a specified
confidence level. Multi-period models can be defined as stochastic models.
Early studies on portfolio selection developed using variance as a risk
measure; although, theories and practices revealed that variance,
considering its downsides, is not a desirable risk measure. To increase
accuracy and overcoming negative aspects of variance, downside risk measures
like semivarinace, downside beta covariance, value at risk and conditional
value at risk was other risk measures that replaced in models. These risk
measures all have advantages over variance and previous works using these
parameters have shown improvements in the best portfolio selection. Purposed
models are solved using genetic algorithm and for the topic completion,
numerical example and plots to measure the performance of model in four
dimensions are provided.
Portfolio selection problem with Value-at-Risk constraints under non-extensive statistical mechanics
