Environmental Engineering Reference
In-Depth Information
same techniques can be extended to more complex, less immediately understood sit-
uations. In addition to more players or decisions, there are possible differences in in-
formation availability and risk aversion that can impact outcomes and decisions.
Information
Information plays a central role in game theory for identifying and achieving pre-
ferred outcomes. The lack of specific pieces of information presents challenges as
well as explanations for suboptimal behaviors and outcomes. A situation with “per-
fect information” means that all possible knowledge about outcomes and prefer-
ences is available to all parties. This state includes “common knowledge” among par-
ties, whereby they all know what each other knows, and know each other has this
information.
The simplest deviation from perfect information is imperfect information. Imper-
fect information describes the situation where the outcome that will result from a par-
ticular set of decisions is not known with complete certainty. Examples include un-
certainty as to whether a particular streambank revegetation project will or will not
increase flooding, or whether a transition from exotic to native plant species will still
attract a certain bird species. If there is disagreement about the most likely outcome
for a particular set of decisions, conflicts can arise. For example, if neighboring farm-
ers expect restoration projects to generate weeds but restoration planners do not, dis-
cussions or additional safeguards might be necessary to alleviate concerns.
The situation is considered one of incomplete information if one party does not
know the preferences or intentions of another party. If residents in an area think that a
particular restoration project is actually intended to capture water rights or crowd out
a particular land use rather than to restore habitat, unnecessary conflict and opposi-
tion might arise. Again, in many cases, simply identifying that certain information is
lacking can identify low-cost solutions.
The approaches to situations of incomplete and imperfect information are some-
what similar and begin with assignment of probabilities to possible scenarios. This is
done by adding branches to the decision tree, probabilistically weighting trees, and
conducting backward induction as before. An example shown in figure 17.2 demon-
strates a situation where a neighboring farmer expects that by supporting Plan B, there
is a 50 percent probability of seeing an increase in revenue (e.g., pest control benefits)
and a 50 percent probability of seeing a decrease in revenue.
If stakeholders have incomplete information about outcomes and restoration
goals, alleviating those uncertainties can prevent probabilistic estimations that would
lead to undesirable behaviors. Identifying situations where stakeholders fear undesir-
able outcomes, even when science suggests such outcomes are unlikely, provide im-
portant opportunities for targeted education and outreach. Or if science is inconclu-
sive regarding the effect, the process can help identify important targets for research
that can yield tangible impacts on project outcomes via behavior changes and coop-
eration. For example, on the Sacramento River farmers were concerned about voles
