Biomedical Engineering Reference
In-Depth Information
Figure 9. SO-Grid graphical interface: Web form that allows the user to compare results related to dif-
ferent simulations
simulation, otherwise this will terminate when the
simulation time expires; in both cases the results
are automatically saved in the file whose name
was specified in the parameter setting phase.
The user can also configure and visualize
a plot that compares results related to different
simulations. Through the form shown in Figure 9,
he/she can check the results files corresponding to
past simulations (in this case results obtained with
different numbers of classes), and the performance
index of interest. After clicking on the “Create
graphic” button, the user will be shown a plot that
compares the values of the chosen performance
index obtained with the selected simulations.
This is a very user friendly and powerful tool
that allows users to carry out “parameter sweep”
analysis of the bio-inspired protocols presented in
this chapter. For example, Figure 10 depicts the
values of the overall entropy obtained in three
simulations in which the number of classes is
set to 3, 5 and 7.
The ant-inspired ARMAP protocol allows for
the dissemination and reorganization of resource
descriptors according to the characteristics of
related resources. This in turn enables the use
of another ant-inspired protocol, ARDIP, which
copes with the discovery of resources on behalf
of users. ARMAP is executed by a number of
ant-like mobile agents that travel the Grid through
P2P interconnections among hosts. Agents dis-
seminate metadata documents on the Grid and
aggregate information related to similar resources
in neighbor Grid nodes, so contributing to decrease
the overall system entropy. Resource replication
and reorganization can be tuned by appropriately
setting a pheromone threshold in order to foster
or reduce the activeness of agents. Simulation
results show that the ARMAP protocol is able to
achieve the mentioned objectives, and is inherently
scalable, as agent operations are driven by self-
organization and fully decentralized mechanisms,
and no information is required about the global
state of the system.
To foster the understanding and use of bio-
inspired algorithms in distributed systems, this
chapter also illustrates the SO-Grid Portal, a
simulation portal through which registered users
can simulate and experience the behaviour of
CONCLUSION AND FUTURE WORk
This chapter proposes a bio-inspired approach for
the construction of a Grid information system.
Search WWH ::
Custom Search