Q-lrnng s rnfrcmnt lrnng lgrthm tht trns n gnt t ssgn vls t ts pssbl ctns bsd n ts crrnt stt, wtht rqrng mdl f th nvrnmnt (mdl-fr). t cn hndl prblms wth stchstc trnstns nd rwrds wtht rqrng dpttns.
Fr xmpl, n grd mz, n gnt lrns t rch n xt wrth 10 pnts. t jnctn, Q-lrnng mght ssgn hghr vl t mvng rght thn lft f rght gts t th xt fstr, mprvng ths chc by tryng bth drctns vr tm.
Fr ny fnt Mrkv dcsn prcss, Q-lrnng fnds n ptml plcy n th sns f mxmzng th xpctd vl f th ttl rwrd vr ny nd ll sccssv stps, strtng frm th crrnt stt. Q-lrnng cn dntfy n ptml ctn-slctn plcy fr ny gvn fnt Mrkv dcsn prcss, gvn nfnt xplrtn tm nd prtly rndm plcy.
"Q" rfrs t th fnctn tht th lgrthm cmpts: th xpctd rwrd—tht s, th qlty—f n ctn tkn n gvn stt.