Vallati, Mauro, Chrpa, Lukáš and Serina, Ivan (2017) On the Evolution of Planner-Specific Macro Sets. In: AI*AI 2017 Advances in Artificial Intelligence : International Conference of the Italian Association for Artificial Intelligence. Lecture Notes in Artificial Intelligence (LNAI) (10640). Springer, pp. 443-454. ISBN 9783319701684

In Automated Planning, generating macro-operators (macros) is a well-known reformulation approach that is used to speed-up the planning process. Most of the macro generation techniques aim for using the same set of generated macros on every problem instance of a given domain. This limits the usefulness of macros in scenarios where the environment and thus the structure of instances is dynamic, such as in real-world applications. Moreover, despite the wide availability of parallel processing units, there is a lack of approaches that can take advantage of multiple parallel cores, while exploiting macros.
In this paper we propose the Macro sets Evolution (MEvo) approach. MEvo has been designed for overcoming the aforementioned issues by exploiting multiple cores for combining promising macros –taken from a
given pool– in different sets, while solving continuous streams of problem instances. Our empirical study, involving 5 state-of-the-art planning engines and a large number of planning instances, demonstrates the effectiveness
of the proposed MEvo approach

evolution-planner-specific.pdf - Accepted Version
Restricted to Repository staff only

Download (235kB)
Add to AnyAdd to TwitterAdd to FacebookAdd to LinkedinAdd to PinterestAdd to Email