Design and Solution of a Surrogate Model for Portfolio Optimization Based on Project Ranking

Table 1

Optimization models for PSPSOP based on proxy variables.

Research work	Description
Fernandez et al. [14]	To the best of our knowledge, this is the first published article that employs a set of proxy variables to solve PSPSOP. The model proposed works on discrepancies , and cardinality of portfolios ; it is as follows:
	.
Fernandez and Olmedo [11]	This work also uses the cardinality of the portfolio but it extends the types of discrepancies used in [14] to include the set of absolute discrepancies , which considers the absolute difference between the considered categories. The model is as follows:
	.
Bastiani et al. [9]	This work improves the multicriteria description of the portfolio qualities, provided by approaches such as [11, 14], through the incorporation of cardinality and discrepancy information per category. Additionally, it includes the power score , an indicator of the number of high-rank projects in a portfolio. The proposed model is as follows:
	,
	where the power and cardinalities are maximized and the discrepancies are minimized.