Multicriteria Portfolio Decision Analysis for Project Selection

Abstract

Multicriteria Portfolio Analysis spans several methods which typically employ build on MCDA to guide the selection of a subset (i.e., portfolio) of available objects, with the aim of maximising the performance of the resulting portfolio with regard to multiple criteria, subject to the requirement that the selected portfolio does not consume of resources consumed by the portfolio does not exceed the availability of resources and, moreover, satis…es other relevant constraints as well. In this chapter, we present a formal model of this selection problem and describe how this model can present both challenges (e.g. portfolio value may, due to the interactions of elements, depend on project-level decisions in complex and non-additive ways) and opportunities (e.g. triage rules can be used to focus elicitation on projects which are critical) for value assessment. We also survey the application of Portfolio Decision Analysis in several domains, such as allocation of R&D expenditure, military procurement, prioritisation of healthcare projects, and environment and energy planning, and conclude by outlining possible future research directions.

Optimization of multiple satisfaction levels in portfolio decision analysis

“…We illustrate the proposed method with the following hypothetical decision problem. Let us suppose that a funding council is expected to select a portfolio of research projects to be realized (for a review of applications in this field, see [60]). In particular, we consider an example involving 20 projects evaluated in terms of the following five criteria [79]:…”

Incorporating balance concerns in resource allocation decisions: A bi-criteria modelling approach

“…In this section we mention some noteworthy studies that consider the balance concept in portfolio selection and allocation decisions in an explicit way. We refer the interested reader to [3] for a more detailed discussion on balance in project portfolio selection problems. There is also a broad range of applications in which equity concerns are incorporated into mathematical models, including but not limited to drug allocation [4], HIV prevention funds allocation [5], water allocation [6], bandwidth allocation [7], workload allocation [8], and location-allocation problems in homeland defense [9].…”

“…Let n = 3, and let the input/output distribution over which balance is sought be a(x): a(x) = (a(x) 1 ; a(x) 2 ; a(x) 3 ) and a(x) 1 +a(x) 2 +a(x) 3 = a T . Suppose the balance distribu- 2 and d(x) 3 be the componentwise absolute deviations from the rescaled balance distribution.…”

Trading off health and financial protection benefits with multiobjective optimization

“…Multiobjective optimization differs from traditional optimization in that it seeks to generate not a single "best" solution, but a range of solutions which may be best depending on where the decision maker sets the tradeoff between competing objectives (Ehrgott, 2005). This can be particularly useful in a portfolio choice setting such as the determination of a benefits package, where making choices "locally" about particular technologies using different criteria may result in global choices which are not logically consistent (Morton, Keisler, & Salo, 2016a;Salo, Keisler, & Morton, 2011).…”