Transcript Multi-Criteria Decision Making and Climate Change

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.[23]
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.