Information Technology Reference
In-Depth Information
- Delay and conflict detection requires the evaluation of the current time-
table (containing all adjusted trajectories) with respect to occupation
overlaps within defined tolerance and time horizon ranges, the detection
of train connection failure risks, the analysis of possible employee or train
vehicle circulation etc. Whatever is evaluated and considered as a problem
respectively a conflict strongly depends on the real implementation and
on the kind of information available. Moreover the dispatching algorithm
also influences the implementation of the conflict detection component.
- The conflict solution component has to solve the identified conflicts and
problems. The algorithm realizing this solving may vary and follow differ-
ent approaches like synchronous or asynchronous paradigms, rule based,
functional or heuristic approaches and much more. As mentioned before
the realization of the dispatching algorithm is the most common sub-
ject of research activities in the context of dispatching systems, including
theoretical consideration with respect to optimality of found dispatching
decisions, the performance, complexity or scalability.
- Dispatching decisions must be integrated within the prognosis on the
one hand and - to make the prognosis probable - propagated into the
real operation on the other hand. The propagation again is varying and
depending on the practical conditions under which the system is work-
ing and the grade of integration within existing systems. A very strong
integration would be the direct influence of operation systems, the di-
rect transmission to train drivers or engines, the information of train
dispatchers, service staff or passenger information systems. A weak inte-
gration would be a simple display for a few people, which have to decide,
what to do with the displayed dispatching system decisions.
Figure 2 shows the different activities and their interactions, which have
to be covered by components of the system architecture. The activities can be
performed by single components, the major interaction and communication
flow between the activities correspond to the once of the components. The
parallelism and scalability of dispatching system components can be derived
from the parallelism of the activities. In the following, activities are often
denoted as (system) components implementing them in a 1:1-relationship.
Components receiving messages can be set up for each communication
channel delivering messages. These components are independent from each
other and act like buffers between operation and the dispatching system. The
components also uncouples the operation time (real time) from the dispatch-
ing system time. It can deliver messages to system internal components on
request, not on a predefined time behavior.
Components processing system messages can be scheduled in parallel ac-
cording to the kind of message. In principle each train and its trajectory
can be handled by a component, train related messages can be assigned to
and evaluated by the component associated to the train. To enrich trajec-
tory information interlocking information can be requested from components
handling interlocking messages. If routing information are available, e.g. al-
