Multi-criteria decision making in climate change planning Patrick Driscoll 5 October 2011 Objectives for today • Reflect a bit on the game play from yesterday • Learn something about decision theory and multi-criteria decision making • Explore a new planning tool with a critical eye Schedule for today • 9-9.50 Review from yesterday and a brief introduction to decision making theory • 9.50-10.00 Break • 10-11 Introduction to Multi-Criteria Decision Making and 1000Minds • 11-12 Group work (in Fib 13, Room 51) • 12-13 Lunch • 13-14.30 1000Minds (in your group rooms, but Martin and I will be available for questions) • 14.30-15.00 Re-group in Room 51 for short group presentations and reflections/comments on the workshop Reflections on the game • What did you learn? • What were some surprising outcomes? • What could we do better next time? Decision and risk • Decision theory and analysis is primarily concerned with how we decide to act in the absence of full information about the consequences of our decisions. • Additionally, we never can be quite sure how alternate pathways may have turned out, but we have to choose. How do we do it? Risk, Uncertainty/Ignorance • Risk is connected to the notion of fortune or chance. • In decision theory, risk is primarily used to indicate the effect of uncertainty on a given objective. The key point is that in this context, risk is needs to be calculable, usually in the form of an expected utility. • Risk = (the probability of the event) X (the expected loss/gain from the event) • When risk is no longer calculable, then we operate in epistemic darkness, BUT we still make decisions all the time. Decision theory (in brief) • Pascal’s choice under uncertainty • Rational actor theory (utility maximization) • Satisficing (the search for adequacy); Herbert Simon, 1956 • Garbage Can Model; Cohen, March, and Olsen (1972) • Prospect theory; Daniel Kahneman and Amos Tversky, 1979 Rational actors • Based on the assumption that we individually seek to maximize our benefits and minimize our costs (at the individual level, not at the aggregate) • Also assumes transitivity. In other words, if you prefer option A to B and B to C, then you prefer A to C. This assumption is a big one, especially in light of presumed perfect information available to the decision taker. Issues with Rational Actor Theory • Embedded assumption that we make decisions based upon dispassionate considerations of utility. • The model suffers from a lack of empirical evidence to support it’s conclusions. • The theory also presupposes that the collective is merely an aggregate of individual decisions. • Both bounded rationality and prospect theory pose significant challenges to the theory of rational actors. Bounded Rationality • Herbert Simon proposed that since perfect information is never available, decision makers rely on “cheats” and shortcuts to arrive at a set of plausible options, then choose amongst this reduced universe. • Rather than seeking to optimize their decisions, decision makers are looking to create outcomes that are satisficient. • Both time and resource limitations constrict the ability of people to rationally weigh all of their options Prospect Theory • Deals with alternatives where the probabilities of alternate outcomes are known or at least calculable • Losses are more important than gains (sunk costs) • Relative changes in utility are more important than absolute changes (we measure against the proximal, not the distal) • Estimation of probabilities are limited by anchoring effects (the first decision anchors all subsequent decisions) • Heuristics (tools) and biases (social, memory, cultural, cognitive, beliefs) • Model-based reasoning also impacts how we decide to act Break (08.50-09.00) Why Use Multi-Criteria Decision Making • Traditional Cost-Benefit Analysis (CBA) is often too narrow to capture non-monetized considerations • Strategic Environmental Assessment and Environmental Impact Assessment tools are often deployed late in the planning stages • MCDM may allow for a broader range of alternative policy decisions to be tested and ranked Examples • What are acceptable losses of biodiversity in infrastructure planning projects • How to value species extinction, degradation of water quality, loss of cultural heritage or permanent land use changes The Alphabet Soup of Decision Making 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. Aggregated Indices Randomization Method (AIRM) Analytic hierarchy process (AHP) Data envelopment analysis (DEA) Dominance-based rough set approach (DRSA) ELECTRE (Outranking) The evidential reasoning approach (ER) Goal programming (GP) Grey relational analysis (GRA) Inner product of vectors (IPV) Measuring Attractiveness by a categorical Based Evaluation Technique(MACBETH) Disaggregation – Aggregation Approaches (UTA*, UTAII, UTADIS) Multi-Attribute Global Inference of Quality (MAGIQ) Multi-attribute utility theory (MAUT) Multi-attribute value theory (MAVT) New Approach to Appraisal (NATA) 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. Nonstructural Fuzzy Decision Support System (NSFDSS) Potentially all pairwise rankings of all possible alternatives (PAPRIKA) PROMETHEE (Outranking) Superiority and inferiority ranking method (SIR method) Value analysis (VA) Value engineering (VE) The VIKOR method. Fuzzy VIKOR. Weighted product model (WPM) Weighted sum model (WSM) Analytic Hierarchy Process • AHP is used to decompose complex problems into a discrete set of sub-problems • For example, climate change can be decomposed into a series of smaller sub-sets of problems, including rising consumption levels, rising primary energy usage, increases in population, large changes in land use, exhaustion of agricultural land necessitating destruction of previously intact forest and rangeland PAPRIKA • Potentially All Pairwise Rankings of all possible Alternatives (yes, I know the “I” is missing somewhere, but I did not create the damn acronym). • Main idea, coupled with conjoint analysis, is that by using the principle of transitivity (if you prefer Option 1 to Option 2, and Option 2 to Option 3, then you prefer Option 1 to Option 3), people can run through the trade-offs between competing goals, visions, measures, desires. Pluses of PAPRIKA • Studies indicate that among all of the different types of multi-attribute/multi-criteria assessment methods, PAPRIKA is the least cognitively demanding. • Proponents claim that the reliability and validity of the choices captured in this method is higher than with other methods (I remain skeptical, but that is part of what today is about) Minuses of PAPRIKA • It is a snapshot in time. There is no guarantee that we will think/feel/perceive the same thing tomorrow or the next day. • The principle of transitivity remains untested in this type of frame. • Once a certain number of variables (over 10) and alternatives are included, the process becomes unwieldy and complex. What does this have to do with climate change? • Good question. Let’s have a look-see. Scalar/Temporal differences “…mitigation and adaptation measures tend to differ in the timing of the efforts (mitigation beneﬁts are lagged in time, unlike some adaptation beneﬁts), the geographical pattern of their effects (mitigation beneﬁts are more global; adaptation beneﬁts are more localized), and the sectoral focus of their responses (mitigation focuses on greenhouse gas emitters and sinks; adaptation focuses on sectors and activities sensitive to climate impacts).” Wilbanks, Leiby, Perlack, Ensminger and Wright (2007), p. 714 Biofuel conversion in Copenhagen Nordhavn (North Harbor), Copenhagen Portland, Oregon 20-minute neighborhood concept Mitigation and Adaptation Synergies Measure Mitigation potential Adaptation potential Coastal wetlands restoration Increased carbon storage Storm buffer, better biodiversity through habitat restoration Sustainable stormwater Lower energy usage management (green roofs, Carbon storage permeable surfaces, passive biological retention) Handle increases in storm and flooding events, increase in habitat within the cities (more biodiversity) Adaptation to Mitigation Conflicts Adaptation measure Adversely affects mitigation Infrastructure relocation from floodplain Higher GHG emissions from new construction Increased urban sprawl due to space limitations Coastal sea wall construction Large increases in energy and materials consumption Negative impacts on erosion, water quality, coastal eco+systems Mitigation to Adaptation Conflicts Mitigation measure Adversely affects adaptation Biofuels Changes in land use, reduction of biodiversity, conflicts with food production, unfairly impacts poorer nations Compact urban development May increase exposure to flooding, storm damage, heat island effects and coastal storm surges Assignment • Use 1000Minds to create a multi-criteria decision model using 4 criteria (you decide) related to climate change planning, rankordered on a 5-point Likert scale (1=least desirable, 5=most desirable). • Then each group prepares a short (5-10 minutes) presentation about what they did and what they found out about their own preferences.