Environmental Engineering Reference
In-Depth Information
Table 18.1
The Prisoner's Dilemma. Two prisoners are held in separate cells for a crime
that they did commit. Each prisoner must decide whether to confess or remain silent about
the crime. The penalties served by each of two prisoners are shown for the possible
combinations of their actions. For example, if neither prisoner confesses the prosecutor
only has sufficient evidence to convict them of a minor offence (1-year imprisonment for
both). However, if one prisoner confesses and the other does not the first goes free and the
second serves a long sentence.
Prisoner B
Confess
Remain Silent
Prisoner A
Confess
5 years for each
0 years for A
20 years for B
Remain Silent
20 years for A
1 year for each
0 years for B
interactions of 'cooperators' and 'defectors', behavioural
economists have shown that altruists (individuals that are
concerned about the interests of others) need not suffer
in material terms because opportunities become available
to them (for example, because they can recognize the
actions of other altruists) that would not be available to
self-serving opportunists.
A second driver underlying the development of
behavioural economics is the recognition that humans
are seldom able to behave like the standard rational
actor assumed by rational choice theory. A pioneer
in the field of artificial intelligence, Herbert Simon
recognized that rarely do actors in the real world
optimize their behaviour, and instead merely try to do
'well enough' to satisfy their goals (Simon, 1957). Simon
named this non-optimal decision-making behaviour
'satisficing', an important component in the concept of
'bounded rationality' that accounts for the limitations
of information, cognitive ability and time that actors
face when making decisions. The presence of bounded
rationality has led some economists to focus on human
reasoning as an inductive process rather than a deductive
one (for example, see Arthurs, 1994). Psychologists have
also found the way in which a problem is framed can
lead to departures from the rational choice model. For
example, Kahneman and Tversky (1981) investigated
an experiment in which people were asked if they
would buy a $10 ticket on arriving at the theatre when
finding themselves in two different situations. In the
first situation they find they have lost $10 on the way to
the theatre and in the second situation they find they
have lost their prepaid $10 ticket. In both situations
the person has lost the value of the ticket ($10) and
therefore, given neoclassical assumptions, should behave
in the same way when considering buying a ticket on
arrival at the theatre. However, Kahneman and Tversky
found that people were overwhelmingly more likely to
buy a ticket in the first situation (88%) than buying a
(replacement) ticket in the second (46%). Kahneman and
Tversky attribute this behaviour to human 'psychological
accounting', in which we mentally allocate resources to
different purposes. In this case, people are less willing
to spend money again on something they have already
debited to their 'entertainment account' than if they have
lost money which they debit to their 'general expenses
account'. We will see below how these behaviours (which
depart from neoclassical models) can be represented in
agent-based models.
18.4 Agent-based modelling
In contrast to neoclassical economics, which examines
the unobserved actions of identical agents (all assumed
to behave in a perfectly rational manner), agent-based
modelling provides a means to represent imperfect, het-
erogeneous actors and their individual decisions, activities
and interactions. These actors can be represented as agents
that make decisions on the basis of individual behavioural
rules and data acquired from their own individual experi-
ences and histories (Tesfatsion, 2002). Agents that adapt
their behaviour according to experience are also possible
via genetic algorithms driven by variation and selection
Search WWH ::
Custom Search