Information Technology Reference
In-Depth Information
developed one such model called SHEBA that required information on a variety of
parameters (i.e., the position on the surface on the Sun from where the event
originated, its time, and its expected speed) to produce reliable resources. These
values are not always known or easily obtained by the users; some may be inferred
from observations, and others can be extracted from existing metadata,
finally and
expert user may be capable determining others simply by their own experience. To
overcome these shortcomings, HELIOGate has developed a more flexible approach
to propagation models that helps users to:
Find the correct values for the execution parameters via assisted propagation
models)
￿
Overcome the dif
finding exact values by executing the models for a
range of values as parameter sweep jobs (Parametric Propagation Models)
culty in
￿
Validate the results of the model by querying metadata catalogues to
nd
signatures of the events on the different planets and satellites via validated
propagation models
￿
￿
Execute the models over multiple events to analyze commonalities among
different phenomena via statistical propagation models
The workflow on the top-right corner of Fig. 14.1 is a simple propagation model
that requires that the parameters of the model be de
ned directly through the input
ports of its only node. Although a simple propagation model can be used on its
own, it is usually the fundamental atomic workflow of more complex models.
The assisted propagation model on the left-top of Fig. 14.1 uses two sets of
nodes to infer some of the parameter values. This is done by invoking a remote
service, an event catalogue, and by downloading an XML
file containing the details
of the event that must be studied. The set of auxiliary nodes, named VOTable
Parsers in the workflow, parses the VOTable returned by the previous nodes and
extracts the relevant data. As an example, the lift-off speed of a coronal mass
ejections is one of the most important parameters of the propagation model, and it is
present in most event catalogues. When a parameter is not present in the catalogue,
it has to be set either to a default value or by the user. A different approach is that of
the validated-parametric propagation model of the bottom of Fig. 14.1 . The vali-
dated-parametric model executes the model for a range of values so that the user has
only to de
ne a range instead of a single value (which is a much easier task). The
results are validated a posteriori by a set of nodes that extract values of the output of
the model and compare them to catalogue of events at the target. An assisted
validated propagation model merges the two approaches by inferring values from
the event catalogues in a fashion similar to the assisted model, but also validates the
results of the execution. Finally, all these models can be executed for multiple
events to infer statistics rather than to investigate a single event. This approach,
which is not shown in Fig. 14.1 , and it is named assisted validated parametric
model.
For reusability reasons, the workflows of Fig. 14.1 are composed of smaller,
atomic workflows. These are query workflows (the catalogue query nodes at the top
and bottom of Fig. 14.1 ) that interface with the metadata catalogues developed by
Search WWH ::




Custom Search