20.500.12556/DKUM-38139 OPTIMIZACIJA DELOVANJA AGENTA V MREŽNEM SVETU S Q-UČENJEM OPTIMISATION OF AGENT OPERATION IN A GRID WORLD USING Q-LEARNING V diplomskem delu predstavljamo samoojačitveno učenje, ki je ena od paradigem učenja. Učenje preučuje optimizacijo obnašanja inteligentnih agentov v okolju, tako da dosežejo čim večjo nagrado. V diplomi opravimo splošen pregled področja, zatem pa implementiramo Q-učenje, ki je ena od metod samoojačitvenega učenja. Delovanje metode demonstriramo na primeru učenja agenta v mrežnem okolju, da preživi čim dlje s pobiranjem koristnih elementov v okolju. In the thesis we present the reinforcement learning, which is one of the paradigms of learning. The learning studies optimisation of the behaviour of the inteligent agents in the environment with the purpose to achieve the best award possible. In the thesis a general field survey is made and after that Q-learning is being implemented Q-learning is one of the methods of the reinforcement learning. The function of the method is shown in the case of an agent in the grid world which has to survive by collecting good kind of elements in the world. strojno učenje samoojačitveno učenje Q-učenje nevronska mreža mrežno okolje machine learning reinforcement learning Q-learning neural network grid world true true false [E. Bibianko] Slovenski jezik Angleški jezik Diplomsko delo/naloga 2012-09-18 10:49:47 2012-12-18 07:50:49 2022-04-25 23:13:23 2021-04-18 03:37:16 2012 0 Maribor 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.85(043.2) 16630806 URN:SI:UM:DK:UHMCDPBQ UNI_Bibianko_Evgen_2012.pdf UNI_Bibianko_Evgen_2012.pdf 1 9224094B4A588D605F5645C62C46D222 e37e6417fc882bb96ada643b6ff8ce4c3ec77511d3702baf5b4d84327371f681 6e93c2d4-7c08-11eb-bb7a-00155d0001ca 20.500.12556/dkum/7bf09b6e-4d2a-464b-83c9-54c767fa2b3a https://dk.um.si/Dokument.php?lang=slv&id=50835 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0