|
|
|
|
|
|
|
Here are some things that
are unique, or at least beneficial, about OpenTS.
|
|
OTS helps you keep focused
on your tabu search by providing a logical one-piece-at-a-time structure for
defining your tabu lists, neighborhood, solution construct, and evaluation
methodology.
|
|
- There are Java classes for
each major element in a tabu search. By defining solution, tabu list, move,
move manager, and objective function classes OTS helps improve development
time by providing a robust engine for grinding through tabu search iterations.
|
|
- OTS enables you to make
reactive, learning, and self-expanding tabu searches without resorting to
"spaghetti code" and without losing the clean design of your tabu
search.
|
|
- OTS can exploit
multiprocessor systems by dividing the neighborhood search among multiple
threads.
|
|
OTS can run embedded in an
Enterprise JavaBeans framework to improve automated decision making.
|
|
|
|
One user who is willing to
be mentioned:
|
|
> half a year ago I
(Fabian Grüning) used OpenTS in my master thesis (computer science) at the
university of Oldenburg, Germany.
|
|
> The starting point is a
given metapopulation model, Meta-X (see
http://www.offis.de/ig/publikationen/publikation.php?id=000668), that
|
|
> can be used to build a
model from a species living in a scattered landscape and is used by
biologists and ecologists. I used OpenTS to
|
|
> automatically provide
allocations of costs to the model's parameters, which correspond to
modifications of the environment in reality, that
|
|
> maximize the species'
viabilty of a modelled landscape and therefore solve an optimization problem.
The calculated allocations can be
|
|
> used by experts for
decision making in landscaping.
|
