In this work we study complexity classes in monotone computation. Our main contributions are the following: ffl A consistent framework for monotone computation, including monotone analogues of many standard computational models. We define monotone simulati

CN1 6 m=NL:Ka chrer mandW gdirsone23] hsoew thatdth uestcon nfncuiot,nw hih cdeciesd

whehtr etwon des aor eoncented in thceg vin undierceet idpunt raph, geruiqes rmnooonetf ormlua ofsde th (pg2l n.) The ylsao itnroudcd ehte ocmmniuactoni copmlxeit lyoewrbo nud ectnihqeu whci ishce tnalr to te prhseentw rk. Sonic uestcon 2n (wLtihnonun fiom arvideci tnhe fom ofr a nuvirsela tarvreas slqeunee c4]) hitsa so lepsaraes mtCN 1 rofm L\ m on.

Moootonend pteh ierahchr: Rya aznd Wgdersoni 1]3s hwodet at hhe matthicn funcgp tio renquies mrnooote nedpth () (none n it thse nuberm f oedesg,)t us prhvongiamon oton edpeht hierrahc yhtoree morf unfctonsi i nP. iSnc meacthign s cioputabml ie nNC R, 2tishi plmei an exsonpentil aag pbetwen monotenoeand onmononotned etp chmolexpiy.

mTC t 06 m=C 1 N: aYo 49] soweh tdht maontonoe oundeb ddpth peolyonima-lsziet herhsodl icrcuis tfrom ap operrhi rearcyh

.t iIs owtr hocntrsaing tthe bova peogrrse sinm ontooen ocpmelxti yithw the cruernt tstae foa iar sn ienegrla ocmpelixyt:as f aras w ekn o, wvere yufction in NnP amyb ec opumetdby edph-thtee trhrshold cireuitsc of olpnyomail siez. orFmor deteialson th ecurret ntats oef ocpmleityxtheo rys ee ht eecretnsu rvyeo Bfoppnaaa dn Spisre 2].11. Ou4 rCntorbituiosnTe sehtitn of ghet respnt eowkr i sshonw ni th eapsmo fuge 1r-1.T h eeftlm ap hsos womotonenco mlepitxyclas ss,ear rnagd vereitaclyl i nrder oof oncainment,t nd athe rigt mhp iast h corerepsondin map ofg egenal (ir.e no.mnnooton)ec asless.N to ehtta lalof thi sccousrwe ll belwo he motre faimlaric alses os fP andNP .Firts w list ehetp eviorslyu nokw neuaqlitisea d snpaearionts o thfsee mpsa.NL=oc-L:N Bythe in dctiuvec onutingtech niuq oef mmerImn a2] 1and Sezelpcesyn i4],2 te clhssa of fucntionsc omutabpe lyb onnedetrinmitisclogs apc euriTg namhcneis is loscde uder nompcelmntaeion.t

51

m