Decision Support

Multi-criteria decision aiding (MCDA) is an integral part of the GaBE approach to informing and supporting decision-makers. GaBE’s emphasis is not necessarily to bring the most cutting-edge theories in MCDA to bear, but rather to match appropriate MCDA methods to the stakeholders’ needs. Some of the key issues include:
  • Ease of use – MCDA methods should be easy both to use and to understand so that stakeholders will trust the results, and interactively interact and learn from the analytic results.
  • Appropriate sophistication – The sophistication of the MCDA theory should match the quality of the data and analytic results.
  • Iterative feedback – Stakeholders’ weights can produce a ranking of technologies or scenarios that contradict their preconceived preferences. This often results in interactive learning and compromise between these prior preferences and their analytic implications.
  • Full integration – The choice of criteria and decision structures should begin before the analysis to make sure that the full range of necessary criteria is produced. Post-MCDA stakeholder learning and feedback can also often lead to better scenario designs for further analysis.
To meet these objectives, GaBE approaches MCDA using two main procedures:
  1. tradeoff analysis
  2. multi-criteria ranking support
Tradeoff Analysis
It is generally impossible to achieve all one’s goals in any constrained problem. The old saying in many situations is “good, fast, cheap – pick any two…” For this reason it is important to understand the tradeoffs in results between the many competing criteria. Pairwise tradeoff analysis allows a graphical presentation of the tradeoffs between any two criteria.

As seen in this schematic figure, the alternatives towards the lower left corner are both cleaner and cheaper. Each point is clearly better than (or dominates) those points in the quadrant above and to the right of it. Together the points in the dominant set form a tradeoff frontier of efficient solutions, so that we only need tradeoffs between these points. The objective is to help individual stakeholders choose their own point on the tradeoff curve, and to learn enough to design better strategies that will move the whole frontier down and left.

This type of analysis has several important strengths. First, the graphic presentation plays to the strengths of the human ability to process information visually and to see patterns. Second, it allows us to discard a large number of inferior alternatives. We can eliminate all the dominated scenarios and only worry about ranking the best alternatives on the tradeoff frontier. Third, if there are many scenarios with sets of competing options, then we can observe patterns in the results and learn a lot about the relative cost and effectiveness of these options.

The main limitation of tradeoff analysis is that it is practically limited to two criteria at a time. A scenario on the tradeoff frontier for one pair of criteria will not necessarily perform well for another pair, and the number of tradeoff graphs that can be remembered and considered is limited. Therefore this approach is most useful where a few pairs of criteria dominate (e.g. cost v. SO2 or CO2), and where we would like to eliminate some of the many scenarios.

Multi-Criteria Ranking Support
When the number of important criteria is large, then the need for decision support requires some form of multi-criteria ranking, as shown in Figure 1.

Dec fig2.jpg

Figure 1

As this figure shows, the ranking process combines stakeholder inputs on the importance or weight of individual criteria with the multi-criteria results for each alternative. An objective function of some kind is then used to reduce the multiple criteria to a single index or score for ranking.

GaBE normally uses weighted averaging as the objective function, for the main reason that it easily allows interactive software that stakeholders can use to adjust their preferences and learn from the changes in the implied rankings. This implies that the necessary steps will include:
  • Making criteria as independent as possible
  • Quantifying qualitative criteria
  • Adjusting criteria to all be “more is better” or “less is better”
  • Transforming data to a common scale (typically 0 to 1 or 0 to 100)
  • Weighting and adding the criteria
If some stakeholders have strong non-linear preferences (i.e. a veto or cut-off response at some level), or a strong risk response, then GaBE has the ability to use other ranking methodologies, but the preference is for easy stakeholder interaction that will allow us to reduce any inconsistencies the stakeholder may have between criteria weights and the ranking results.

Main projects and selected results

New Energy Externalities Developments for Sustainability (NEEDS)
The objective of the four year (2004-2008) European Integrated Project NEEDS (New Energy Externalities Developments for Sustainability) is to “evaluate the full costs and benefits (i.e. direct + external) of energy policies and of future energy systems, both at the level of individual countries and for the enlarged EU as a whole.” To achieve this, the ambition is to integrate three major methodologies: LCA, External Cost Assessment and Energy-Economy Modeling (Markal/Times), as well as apply MCDA (the latter under PSI-GaBE lead). NEEDS is a project of the ExternE Series. Within the MCDA effort, PSI-GaBE is responsible for coordinating selection of sustainability criteria, choice of an appropriate MCDA methodology, implementation of MCDA software, and stakeholder interaction.

China Energy Technology Program (CETP)
The Electric Sector Simulation task of the China Energy Technology Program (CETP) involved detailed dispatch modeling of Shandong province over 25 years. The results from this analysis were integrated with results from other tasks, including life cycle assessment, environmental impacts and external costs, and risk of severe accidents. Results for a wide range of sustainability criteria were used in a decision support process to assist a range of Chinese stakeholder participants who formed the stakeholder advisory group. Their input was used for scenario design and criterion weighting. The ELECTRE III methodology was used to ascending and descending rankings, including criterion vetos. In addition, a weighted sum methodology was used for interactive decision support software that was included as part of the DVD that accompanied the CETP book.

Strategic Electric Sector Assessment Methodology under Sustainability Conditions (SESAMS)
The SESAMS project, short for “Strategic Electric Sector Assessment Methodology under Sustainability Conditions”, was a joint project under the Alliance for Global Sustainability (AGS) between the Paul Scherrer Institute (PSI), the Swiss Federal Institute of Technology in Zurich (ETHZ) and the Massachusetts Institute of Technology (MIT). The project was the first complete integration of the electric sector modeling and life cycle assessment expertise of the partners, and covered three phases between 1997 and 2001. A stakeholder advisory group provided input for issue identification, scenario design, approval of major assumptions, and choice and weighting of criteria for decision support using the ELECTRE methodology.