This paper considers the general problem of relocating modules to convert infeasible constrained floorplanner inputs into feasible ones. This is accomplished by placing modules in locations that attempt to minimize the standard deviation of module densities. Efficient geometric algorithms are developed and shown to be successful in obtaining feasible inputs. Experimental results examine the tradeoffs between achieving a good redistribution of density on one hand and minimizing the increase in

more »

... re length and the displacement of modules on the other. Index Terms-Algorithm, computational geometry, incremental floorplanning, placement, standard deviation. . His research interests include design and analysis of algorithms and computer-aided design (CAD) of very large scale integration (VLSI) physical design. Dinesh P. Mehta (S'90-M'92-SM'04) received the B.