Link Search Menu Expand Document

Cerberus: Red Black Planning with Mutex Detection and Novelty Heuristic

Page Contributors: Adam Green,

tags: IPC2018 Satisfycing

Planner Quality: -

Year Published: 2018

Paper: Cerberus: Red-Black Heuristic for Planning Tasks with Conditional Effects Meets Novelty Heuristic and Enchanced Mutex Detection [ Katz, M. ]

Preceded By: Mercury

Cerberus is built on top of Mercury, a planner which makes use of a Red-Black Heuristic (see our page on Red-Black Heuristics [In Progress]). With the introduction of novelty based heuristic guidance, Cerberus is designed to have better support for Conditional Effects. Furthermore, Cerberus does “Enhanced [Mutual Exclusion] Detection” to derive a SAS+ Planning task (see our page on SAS+ Planning [In Progress]). This allows it to perform a best-first search using the red-black planning heuristic.


Cerberus has not been tested with eviscerator as we could not get the source code to compile in a way to test it. Cerberus appears to be targeted to solving classical planning problems, so likely does not support any temporal or numeric features in PDDL.

Downloading and Compiling Cerberus

Cerberus planner source code with a singularity container support can be found as part of a BitBucket Submission to IPC 2018

Cerberus-gl has source code located here submitted to the same IPC