Task-Driven Species in Evolutionary Robotic Teams - Foundations on Natural and Artificial Computation - page 137

Task-Driven Species in Evolutionary Robotic
Teams
P. Trueba, A. Prieto, P. Caamano, F. Bellas, and R.J. Duro
Integrated Group for Engineering Research,
Universidade da Coruna, 15403, Ferrol, Spain
{pedro.trueba,abprieto,pcsobrino,fran,richard}@udc.es
http://www.gii.udc.es
Abstract. This paper deals with the problem of obtaining coordinated
behavior in multirobot systems by evolution. More specifically, we are in-
terested in using a method that allows the emergence of different species
if they are required by the task, that is, if specialization provides an
advantage in the completion of the task, without the designer having to
predefine the best way to solve it. To this end, in this work we have ap-
plied a co-evolutionary algorithm called ASiCo (Asynchronous Situated
Co-evolution) which is based on an open-ended evolution of the robots
in their environment. In this environment the robots are born, mate and
die throughout the generations as in an artificial life system. In order to
show that ASiCo is capable of obtaining species automatically if they are
advantageous, here we apply it to a collective gathering and construction
task where homogeneous teams are suboptimal.
Keywords: Multi-robot Systems, Evolutionary Algorithms, Coordina-
tion, Collective Intelligence.
1 Previous Work
Using evolution for obtaining coordination in a collective system is not new
[4] and different evolutionary algorithms have been developed and applied for
years [3][5]. An open issue in this field is that of emergent specialization, that is,
the automatic definition of heterogeneous teams in response to a dynamic task
that requires individuals performing different sub-tasks in order to accomplish
the global one. Specialization can be morphological [12] or behavioural [2]. The
first question to answer is which type of task would require specialization. In
this sense, there is some agreement amongst researchers that if the task can be
naturally decomposed into a set of complementary sub-tasks, then specialization
is often beneficial for increasing collective task performance [2]. Examples of
these tasks are collective gathering [7], collective communication [6] or multi-
agent computer games [5].
In the last decade, coevolutionary approaches have provided the most remark-
able results in emergent specialization [13]. Within this line, the authors in [1]
conclude that the best approach consists in evolving each robot controller in a