Information Technology Reference
In-Depth Information
Social Welfare for Automatic Innovation
Juan A. Garca-Pardo and C. Carrascosa
Universitat Politacnica de Valencia
Camino de Vera s/n. 46022 - Valencia (Spain)
{
jgarciapardo,carrasco
}
@dsic.upv.es
Abstract.
Individuals inside a society can make organizational changes
by modifying their behavior. These changes can be guided by the out-
come of the actions of every individual in the society. Should the outcome
be worse than expected, they would innovate to find a better solution to
adapt the society to the new situation automatically.
Following these ideas, a novel social agent model, based on emotions
and social welfare, is proposed in this paper. Also, a learning algorithm
based on this model, as well as a case of study to test its validity, are given.
1
Introduction and Motivation
Humans, as social beings, are situated in a society that influences them. In the
same way, their actions and opinions may influence their society in a higher or
lower degree, depending on their individual position in it; that is, the reputation
and the amount of other people who is aware of their actions and opinions.
Their opinion will have a relative impact in others: the importance depends
on how much their opinion is trusted, and how much of an expert they are.
Different personalities will react differently to the same stimulus, which means
that different persons will be easier to persuade than others by means of social
welfare.
For instance, lets have an individual which is baking bread in the bakery.
This baker only knows about what is visible from its position, e.g. flour prices,
bread production,. . . Should a problem with wheat production affect flour prices
later, it would go unnoticed by the baker, unless the flour providers
express
their
opinion of how well the flour production is going. If the baker takes into account
that opinion, he would start to innovate in the way he produces bread, for
instance, by producing it with less flour, or by baking bagels if they are bought
by his customers. Of course, different bakers could have different personalities,
which could push them towards innovation more easily or not. The aggregated
opinion of the society can be viewed as an indicator of the society's welfare, as
all the individuals express their opinion about their situation. We hypothesize
that this opinion is useful for adaptation purposes, without the necessity of a
change detection mechanism.
Following this idea, in this paper an agent model -based on a well studied
psychological model of emotions- is proposed in section 3. Using this model, an
algorithm is provided in section 4, as an example of how the model could be used;
