Information Technology Reference
In-Depth Information
The social context is represented as a directed social network where a network link
from agent A to agent B means that B is influenced by A. In the model, the invest-
ment behaviour of A is perceivable by B and contributes to B's evaluation of its social
conformity goal along with the behaviours of all other agents having outgoing social
network links to B. In principle, an agent's social network peers may be located on
arbitrary patches of the environmental context, i.e. network links may span agent
groups.
HAPPenInGS forms the theoretical framework of agent decision-making in HAP-
PenInGS-A. Accordingly, an agent's selection of a behavioural option is guided by
the three preference dimensions stated by the theory which are shown in the first col-
umn of Table 1. An agent's knowledge about the effectiveness of his behavioural
options is represented by a utility calculation which relates the perceptions of an agent
and his expectations about the outcome of executing a behaviour to the agent's
subjective preferences.
Table 1.
Preferences according to HAPPenInGS, the respective parameters, and assessment
criteria along with an example preference set. An agent's preferences are weighting factors for
the criteria. The last column displays the formulas used to calculate an agent's estimation of the
satisfaction of a given preference. Here x is the investment level of the behavioural option
evaluated, c(x) stands for the expected level of the public good, abs() calculates the absolute
value.
HAPPenInGS
preference
Preference parameter
(example value)
Criterion
publicGoodPreference
(1.0)
Public good
c(x)
selfPreference
(0.1)
1.0 - x
Distribution of
investments
othersPreference
(0.3)
1.0 - mean x of other
group members
1.0 - abs(x - mean x
of social network peers)
In HAPPenInGS-A, agents use this subjective utility to assess, compare and rank
their behavioural options based on their preferences and perceptions. To do so, each
agent perceives the present success of the collective action, i.e. the level of the public
good provided by its respective group, and supposes that the n-1 other agents of its
group keep to their previous investment decisions in the next time step. Thus, each
agent can estimate the success of the collective action associated with each of its pos-
sible next investment decisions. Moreover, an agent perceives the average level of
contributions within its group which enables the assessment of the expected confor-
mity of an investment behaviour with a given preference for the distribution of in-
vestments in the providing group. Finally, agents refer to their social network and
determine the average level of contributions by their peers as a proxy for a perceived
social norm. Table 1 gives an overview of the preference structure along with an ex-
ample preference set. Based on their preference set, agents can determine the
socialConformityPreference
(0.1)
Social conformity
