OnHandling theFixed-outline Constraints ofFloorplan- ningUsingLessFlexibility First Principles *

evolutionary search tohandle fixed-outline floorplanning based on thenormalized polish expression. Liuetal.[7] proposes anapInthis paper, wepropose anefficient approach onhandlingprahfrixdotnelopanngbsdncmonusthefixed-outline constraints withtheobjectives ofbothimprovingprahfrixdotnelopanngbsdncmonusthefixed-ouatle constraintsiwi the o iveofnbth. iprovin quenceofsequence pair [2]. ChenandChang [8] present anadapthesuccess rateainimizn tire l h.Oraproachtive fast simulated annealing algorithm based onB*atree thatdyusestheLessFlexibility First principles [1]whicharederived nmclycagstewihsi h otfnto ne h fromhuman's experience insolving rectangle packing problem in fix-lin constraints. fixed-outline constraints. thedaily life. Inthis construction basedapproach, success rate is Inthis paper, we introduce aconstruction basedapproach on guaranteed byusing effective heuristics toguide theplacementhandling thefixed-outline constraints using theLessFlexibility process. Forthewire length objective, weuseawirelength mini- First (LFF) principles [1]. Theobjectives aretobothimprove the mized solution gained bycircuit partitioning andmultilevel simu- successrateundertight outlines andminimize thewirelength. lated annealing asa reference, wherethetopological relationsHigh success rate isguaranteed byusing effective heuristics and derived fromitaretried tobekept. Wealso apply alocal groupingtechniques toguide theplacement process. Forthewirelength strategy andtheminimal local wirelength objective assupple- objective, we useawirelength minimized solution asareference mentsofwirelength optimization. Experiment results showthat [9], wherethetopological relations derived fromit aretried tobe ourapproach canachieve high successrates undertight outlineskept. We address thatsuchareference solution canbequickly andcomparable wirelength withvarious aspect ratios. generated byconnectivity basedcircuit partitioning andmultilevel simulated annealing. Further, weapply alocal grouping strategy