Kwouo1ne8 ulwjea‘uoI4eo1dxa←jeuoney Injssa3ns卡jeuole!‘3uH panoadxe se aq qou Kew sawooino uonoe- ueKomiepp卡jeuoney uonewoju!quenaja le Kjddns qou Kew sidaouad- 1uB1 sIuw0卡Ieuo14eH anep o1 aouanbas ndaoJad ayl uanl8 anseaw aouewojad ay Jo anjen panpadxa ay1 saziwixew uonpe JanayplyM sasooyo quage jeuoney is3enbsμIpY<oy3z到euad- ianow Jad auo snuiw 'dans awn Jed auenbs ueajo jed quiod auo- iL aw!u!dn paueaj auenbs jed quiod auo- eouanbas quawuollnua ay1 sanenjena aunseaw aouewjojad paxl AuleuoleH y Rationalit sequence environment the evaluates measure rmance erfo p Fixed ? T time in up cleaned re squa er p oint p one – move? er p one minus step, time er p re squa clean er p oint p one – res? squa y dirt k >r fo enalize p – of value ected exp the maximizes action whichever oses cho agent rational A date to sequence ercept p the given measure rmance erfo p the omniscient 6= Rational rmation info relevant all supply not y ma ercepts p – ant y clairvo 6= Rational ected exp as eb not y ma outcomes action – successful 6= rational Hence, autonomy rning, lea ration, explo ⇒ Rational 7 2 Chapter