Risk-Based ROI, Capital Budgeting, and Portfolio Optimization in the Department of Defense https://www.dau.edu
Alternative Analytical Approaches This section delves into some alternative analytical approaches that can be used to determine an optimal portfolio. This discussion is included to provide the reader and decision maker a better sense of what else lies beyond the realm of portfolio optimization, as well as some basic technical aspects and their pros and cons. Finally, this section suggests opportunities for future research. Lexicographic Average Rank for Evaluating Uncertain Multi- Indicator Matrices with Risk Metrics In many situations, projects are characterized by several criteria or attributes that can be assessed from multiple perspectives (financial, economic, etc.). Each criterion is quantified via performance values, which can either be numerical or categorical. This information is typically structured in a multi-indicator matrix Q . A typical problem faced by a decision maker is to defne an aggregate quality (AQ) able to synthesize the global characteristics of each project and then derive the rankings from the best to the worst base-case ranking (Mun et al., 2016). Ranking techniques can be classifed as parametric and nonparametric. Aparametric technique requires information about decision-maker preferences (e.g., criterionweights). According to Dorini, Kapelan, and Azapagic (2011), some examples of parametric techniques include the ELECTRE (Elimination Et Choix Traduisant la Realité or Elimination and Choice Expressing Reality) methods (Roy, 1968) and PROMETHEE—Preference Ranking OrganizationMethods for Enrichment Evaluations (Brans &Vincke, 1985). Nonparametric techniques, such as Partial Order Ranking (Bruggemann, Bücherl, Pudenz, & Steinberg, 1999) and Copeland Scores (Al-Sharrah, 2010), do not require information from the decision maker. In general, all of these techniques are able to produce a ranking of the alternatives from the best to the worst. Therefore, given a matrix Q , the selected procedure generates a ranking, defined as the base-case rank (BCR). As a result of this assessment, for each alternative, a specifc rank R i that considers the multiple perspectives defned by the decision maker is obtained. The set of R i corresponds to the global evaluation under the frst synthetic attribute, defned and named as base ranking , and capable of characterizing the alternatives in the base case. However, each performance value could be afected by uncertain factors. Several approaches have been presented for analyzing how the uncertainty in the performance values (the input) affects the ranking of the objects
78
Defense ARJ, January 2020, Vol. 27No. 1 : 60-107
Made with FlippingBook Learn more on our blog