Operations Research 3.0 Adds Reinforcement Learning and
Improved Shortest Path Algorithms
January 17, 2003--The new Operations Research 3.0 application
provides business professionals and students with even more tools for
solving problems in linear optimization, quadratic programming, and
combinatorial optimization--with heuristic routines for the hardest of these
instances. The updated product also offers new methods for shortest path
tasks and reinforcement learning.
Operations Research consists of five component packages, one of
completely new in this version. It addresses the topic of reinforcement
learning, a relatively new method for stochastic optimization and control
that is quite powerful for certain types of problems. The package on
shortest path tasks has also been completely rewritten for Version 3.0. It
includes a new, very efficient, point-to-point routing algorithm as well
as an improved implementation of Dijkstra's algorithm.
The chapter on combinatorial optimization has been enlarged in Version 3.0
by implementing more heuristics, notably tabu search, while the
comprehensive documentation and selection of example notebooks has been
expanded. Each problem is explained, modeled, and solved in full detail.
For more information, refer to the Operations Research