Evaluation of real-time search algorithms in dynamic worlds (summary of results)

S. Shekhar, B. Hamidzadeh

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations

Abstract

There are many application domains of real-time Artificial Intelligence (AI), where the world changes during the problem solving process. Several real-time search algorithms have been proposed for problem solving in dynamic environments. However, there has not been any systematic evaluation and comparison of these algorithms. This paper provides a classification of different dynamic worlds. It then provides a detailed model of a dynamic world where changes occur in edge costs around a zero mean. A formal analysis of the model suggests that the static rank ordering of solution paths is preserved in the proposed dynamic model. The paper provides analysis of two real-time search algorithms, namely DYNORAII and RTA∗, for the real-time path planning problem. We provide new results on the path planning problem in the proposed dynamic model of graphs. We also provide experimental evaluation of DYNORAII and RTA∗ in their ability to minimize response-times in dynamic environments.

Original languageEnglish (US)
Title of host publication4th International Conference on Tools with Artificial Intelligence, ICTAI 1992
PublisherIEEE Computer Society
Pages6-13
Number of pages8
ISBN (Electronic)0818629053
DOIs
StatePublished - Jan 1 1992
Event4th International Conference on Tools with Artificial Intelligence, ICTAI 1992 - Arlington, United States
Duration: Nov 10 1992Nov 13 1992

Publication series

NameProceedings - International Conference on Tools with Artificial Intelligence, ICTAI
Volume1992-November
ISSN (Print)1082-3409

Conference

Conference4th International Conference on Tools with Artificial Intelligence, ICTAI 1992
CountryUnited States
CityArlington
Period11/10/9211/13/92

Fingerprint Dive into the research topics of 'Evaluation of real-time search algorithms in dynamic worlds (summary of results)'. Together they form a unique fingerprint.

Cite this