EXCALIBUR
Adaptive Constraint-Based Agents in Artificial Environments

[ORC QUEST REVISITED]   [Constraints]   [Heuristics]   [Global Search Control]   [Evaluation]

[ Please note: The project has been discontinued as of May 31, 2005 and is superseded by the projects of the ii Labs. There won't be further updates to these pages. ]

The Global Search Control

(Related publications: [PUBLink] [PUBLink])

Within the global search control, the Performers constraint will always be chosen to perform an improvement if the plan does not yield at least 5 performers. Otherwise, if the plan yields any pain, the Pain constraint will be chosen. If we have a plan with enough performers and no pain, the Orc constraint will be chosen to help shorten the plan. This enables a search to be realized that has the performers criterion as its satisfaction goal, the pain criterion as the primary optimization goal and the duration criterion as the secondary optimization goal.

The search process is initialized with a CSP that includes the Performers, Pain and Orc constraint and a Task Constraint with an ActionType variable that represents the task constraint's fourth configuration (to establish the satisfaction goal Performers >= 5). The state resource constraints Performers and Pain are initialized with CurrentState values of 0.


[ORC QUEST REVISITED]   [Constraints]   [Heuristics]   [Global Search Control]   [Evaluation]

For questions, comments or suggestions, please contact us.

Last update:
May 20, 2001 by Alexander Nareyek