Rectangle-Packing-Based Module Placement

Murata, Hiroshi; Fujiyoshi, Kunihiro; Nakatake, Shigetoshi; Kajitani, Yoji

doi:10.1007/978-1-4615-0292-0_42

Cited by 85 publications

(136 citation statements)

References 5 publications

Supporting

Mentioning

136

Contrasting

Order By: Relevance

“…Murata et al (1996) introduced an elegant P-admissible packing representation called Sequence Pair (SP), which became widely accepted in both academia and industry societies. The certain confirmation of this fact, is that the work (Murata et al 1995) has been selected as one of the 40 best papers published at the IEEE/ACM International Conference on Computer Aided Design (ICCAD) during the past 20 years (Kuehlmann 2003).…”

Section: Block Packing Problemmentioning

confidence: 90%

New perspectives in VLSI design automation: deterministic packing by Sequence Pair

2008

View full text Add to dashboard Cite

In the paper we consider a problem of packing rectangular blocks on a plane, which is known as Block Packing Problem (BPP). This problem is a central issue of the modern VLSI chips design methods. Basing on a new interpretation of the Sequence-Pair representation of the packing solution-space, which is based on Complementary Mirror Constraint Graphs (CMCG), we develop the efficient method of neighborhood exploration. This method might be able to improve the efficiency of other neighborhood-based search methods, such as simulated annealing or tabu search, as well as, to construct efficient heuristics. We illustrate application of the developed method by constructing a heuristic algorithm for solving BPP and comparing its efficiency and effectiveness to the algorithms commonly used so far.

show abstract

Section: Block Packing Problemmentioning

confidence: 90%

New perspectives in VLSI design automation: deterministic packing by Sequence Pair

2008

View full text Add to dashboard Cite

show abstract

“…The vast majority of studies have focused on generating approximate or suboptimal solutions, using techniques such as genetic algorithms and simulated annealing. To facilitate these approaches, a number of data structures have been developed to represent a layout of rectangles including sequence pairs (Murata et al 1995), BSG structures (Nakatake et al 1996), and O-Trees (Guo et al 1999). We omit the details of such structures, as they are largely unrelated to our formulations.…”

Section: Previous Work On Optimal Rectangle Packingmentioning

confidence: 99%

Optimal rectangle packing

Korf

Huang

2008

Ann Oper Res

View full text Add to dashboard Cite

We consider the NP-complete problem of finding an enclosing rectangle of minimum area that will contain a given a set of rectangles. We present two different constraintsatisfaction formulations of this problem. The first searches a space of absolute placements of rectangles in the enclosing rectangle, while the other searches a space of relative placements between pairs of rectangles. Both approaches dramatically outperform previous approaches to optimal rectangle packing. For problems where the rectangle dimensions have low precision, such as small integers, absolute placement is generally more efficient, whereas for rectangles with high-precision dimensions, relative placement will be more effective. In two sets of experiments, we find both the smallest rectangles and squares that can contain the set of squares of size 1 × 1, 2 × 2, . . . , N × N , for N up to 27. In addition, we solve an open problem dating to 1966, concerning packing the set of consecutive squares up to 24 × 24 in a square of size 70 × 70. Finally, we find the smallest enclosing rectangles that can contain a set of unoriented rectangles of size 1 × 2, 2 × 3, 3 × 4, . . . , N × (N + 1), for N up to 25.

show abstract

“…As a fundamental geometric problem, it has been studied for many decades [33]. Algorithmic advances were made in the context of VLSI design, where arrangements of rectangles are represented with certain structures, for example sequence pairs [34], bounded sliceline grids [35], Q-sequences [50] or B*-trees [41], that can be improved by heuristics like simulated annealing. To compute wirelength optimal placements a floorplan representation is needed that allows to represent placements of floorplans in general position.…”

Section: Related Workmentioning

confidence: 99%