Speaker
Describe the added value of the Grid for the scientific/technical activity you (plan to) do on the Grid. This should include the scale of the activity and of the potential user community and the relevance for other scientific or business applications
The main problem is that a mass of data acquired from large
experiments could hardly
be processed. System identification and generalization of
observations problems are
well-known in data mining. When the studied processes are complex
and involve a great
number of variables, solving this task by the ordinary means is
too time-consuming.
Grid infrastructure enables an implementation of methods for the
conversion of
multidimensional raw data into useful information. It seems
encouraging for the
technical domains where reliable real-time prediction and control
is critical.
With a forward look to future evolution, discuss the issues you have encountered (or that you expect) in using the EGEE infrastructure. Wherever possible, point out the experience limitations (both in terms of existing services or missing functionality)
- It would be nice to have some automatic regular middleware
means for dynamical
scheduling of new jobs based on the results of already completed
ones. A good example
of dynamical finding of good-enough solution is the computer chess. - Native storages representation. There are many experimental
non-grid databases
using a variety of data formats. While solving some task only a
small part of the
data is usually needed, so it is unpractical to transfer all the
data into storage
and computing e
Report on the experience (or the proposed activity). It would be very important to mention key services which are essential for the success of your activity on the EGEE infrastructure.
To solve these multipurpose optimization problems the authors
have exploited some
kind of genetic algorithm (GA). GA implies a large number of
independent
calculations, so it can be easily parallelized by means of grid
computing.
A testing model of plasma transport inside ITER has been found.
Plasma transport has
been simulated by ASTRA code under various physical conditions
(initial plasma state,
controlled plasma heating) in 100,000 runs. Then, the optimal
structure of ordinary
differential equations (ODE) describing plasma transport
depending on external
influence and plasma state has been derived. The ASTRA
simulations and the
optimization were performed by GA method using Russian Data
Intensive Grid
infrastructure.
To implement the Grid computing technique heterogeneous software
has been glued by
Python and Shell scripts. Though such an approach is flexible
enough, it becomes
rather tangled with the growth of number of software components.
Describe the scientific/technical community and the scientific/technical activity using (planning to use) the EGEE infrastructure. A high-level description is needed (neither a detailed specialist report nor a list of references).
The activity aims to offer a method for extraction of the most
significant dynamical
input-output dependencies from the large data sets using the
Grid. This work has
mainly been done to generalize cumulated observations in the
nuclear fusion science
for the purpose of model-based controllers improvement, though an
approach can be
employed for any domain where a short-range prediction is needed.
