Descripción
|
|
---|---|
Asaresultofterroristattacksinthelastyears,neweffortshaveraisedtryingtosolvechallengesrelatedtose- curitytaskautomationusingroboticplatforms.Inthispaperwepresenttheresultsofacooperativemulti-robot approachforinfrastructuresecurityapplicationsatcriticalfacilities.WeformulateourproblemusingaMs. Pac-Maclikeenvironment.Inthisimplementation,multipleroboticagentsdefinepolicieswiththeobjectiveto increasethenumberofexploredstatesinagridworld.Thisisthroughtheapplicationoftheoff-policylearning algorithmfromreinforcementlearningarea,knownas Q-learning. Wevalidateexperimentallyourapproach withagroupofagentslearningapatroltaskandwepresentresultsobtainedinsimulatedenvironments. | |
Internacional
|
Si |
Nombre congreso
|
ICAART 2012 - International Conference on Agents and Artificial Intelligence |
Tipo de participación
|
960 |
Lugar del congreso
|
Portugal |
Revisores
|
Si |
ISBN o ISSN
|
978-98-9842-596-6 |
DOI
|
10.5220/0003721403130316 |
Fecha inicio congreso
|
06/02/2012 |
Fecha fin congreso
|
08/02/2012 |
Desde la página
|
313 |
Hasta la página
|
316 |
Título de las actas
|
Proceedings of the 4th International Conference on Agents and Artificial Intelligence |