| << Prev | - Up - | Next >> |
This chapter briefly outlines the features of the Oz Explorer.
Invoking
The Explorer is invoked by applying it to a script. It can be applied to both a script and an order, which switches the Explorer to branch and bound mode such that it supports best solution search.
Exploring
The search tree of the current script can be explored in a user-guided manner. Starting from any node in the search tree further parts can be explored. Explored parts of the search tree are drawn as they are explored.
Search trees can be explored in a depth-first fashion up to the next solution. Entire subtrees as well as single nodes can be explored. Exploration and/or drawing can be stopped and resumed at any time.
Information Access
Nodes in the tree representing choices and solutions carry as information their corresponding computation spaces. The Explorer gives first-class access to them by predefined or user-defined procedures.
Statistical information is available in a status bar during exploration. Detailed statistical information with respect to a subtree can be accessed by predefined or user-defined procedures.
Display Economics
The display of large search trees can be kept economic by scaling the tree and by hiding subtrees. Support for automatically hiding complete subtrees not containing solutions is provided. Hidden subtrees can be unhidden on demand.
Recomputation
The Explorer can run scripts with a large number of constraints and propagators and with large search trees efficiently with respect to both space and time. User-configurable recomputation can be employed to trade space for time for scripts which would require to much memory otherwise.
Postscript Output
The search tree of a script can be dumped in postscript format.
Finding Nodes
Nodes in the search tree can be selected directly or by convenient short cuts.
| << Prev | - Up - | Next >> |