Selecting a Portfolio of Solar Energy Projects Using Multiattribute Preference Theory

Abstract

This article reports a procedure developed to assist the U.S. Department of Energy in selecting a portfolio of solar energy applications experiments. The procedure has also been used in other government procurements and appears to be applicable in a variety of project funding processes. The technical quality of each proposed applications experiment was summarized through the use of multiple evaluation measures, or attributes. These were combined into a single index of the overall technical quality of an experiment through the use of a multiattribute utility function. Recently derived results in measurable value theory were applied to derive an index of the overall technical quality of a portfolio of experiments. Budgetary and programmatic issues were handled through the use of constraints. This approach allowed the portfolio selection problem to be formulated as an integer linear program. Details of the application are presented, including a discussion of the data requirements and assessment procedure used. The portfolio selection procedure was successfully applied, and variations of it have been successfully used in four other solar energy procurements.research and development: project selection, utility/preference: multiattribute, programming: integer, applications

Use of Decision Analysis in the Army Base Realignment and Closure (BRAC) 2005 Military Value Analysis

“…Parnell (2006) compares 10 single-decision applications (similar to our installation model) and 14 portfolio decision value model (similar to our portfolio model) applications. Additional portfolio decision models include Golabi et al (1981), Archer and Ghasemzadeh (1999), Kleinmuntz and Kleinmuntz (1999), and Stummer and Heidenberger (2003 We reviewed army, DoD, and joint service militaryrelated documents and published reviews (e.g., Government Accountability Office, RAND Corporation, etc. ), focusing on defense transformation, stationing, and BRAC.…”

“…After questioning the Army Basing Study installation experts, it was fairly straightforward to Golabi et al (1981), we are able to make simplifying assumptions for the assessment process by decomposing the set of attributes for a particular installation such that each expert only needs to assess the set of M attributes x 1j x 2j…”

Measurable Multiattribute Value Functions for Portfolio Decision Analysis

“…We begin with an illustrative example to introduce the standard framework for multiobjective portfolio decision analysis by Golabi et al (1981). The example is based on pilot projects carried out in Finland in which public funds were used to purchase privately owned forest sites for conservation for a fixed period of some 10-20 years.…”

“…Selecting which portfolio of project candidates (e.g., products, infrastructure investments, research programmes, policy options) to implement with the available resources is an important decision problem both in public administration and private companies (e.g., Golabi et al 1981, Ewing Jr. et al 2006, Liesiö et al 2007, Kleinmuntz 2007, Phillips and Bana e Costa 2007, Grushka-Cockayne et al 2008. Maximizing the value of a project portfolio is often complicated by multiple value criteria and a large number of alternative portfolios.…”

“…Plenty of applied and theoretical research has been done to develop analytic support for such decisions, mainly under titles such as portfolio decision analysis, project portfolio selection and resource allocation (for overviews see, e.g., Klein-constraints can be introduced to limit the sum into a range where the marginal value is approximately constant (e.g., Kleinmuntz 2007). However, these approaches are not directly applicable for 'non-additive' attributes or attributes without a natural measurement scale, which both are common in applications of multiobjective portfolio decision analysis (e.g., Golabi et al 1981, Ewing Jr. et al 2006, Liesiö et al 2007, Kleinmuntz 2007, Phillips and Bana e Costa 2007. As an exception, Grushka-Cockayne et al (2008) use non-additive criterion specific portfolio value functions in the In this paper we develop non-additive portfolio value functions based on the theory of measurable multiattribute value functions by Dyer and Sarin (1979).…”