Matthias F. Stallmann
ABSTRACT
After extensive experiments with two algorithms, CPLEX and our implementation of all-integer dual simplex, we observed extreme differences between the two on a set of design automation benchmarks. In many cases one of the two would find an optimal solution within seconds while the other timed out at one hour.
We conjecture that this contrast is accounted for by the extent to which the constraint matrix can be made block diagonal via row/column permutations. The actual structure of the matrix without the permutations is not important.
We used crossing minimization to discover the right permutations that make the underlying structure, whether block diagonal or random, visible. Our conjecture leads to a testable hypothesis for a limited domain.
The hypothesis is based on earlier observations reported in an earlier version of this paper and subsequent exploratory experiments. We present a sample of the results validating the hypothesis and make additional observations outside its scope.
As far as we are aware the approach taken here is unique and, we hope, will inspire other research of its kind.
- C. Aykanat, A. Pinar, and Ü. V. Çatalyürek. Permuting sparse rectangular matrices into block-diagonal form. SIAM J. Sci. Comput., 25:1860--1879, 2004. Google Scholar
Digital Library
- R. Borndörfer, C. E. Ferreira, and A. Martin. Decomposing matrices into blocks. SIAM J. Optimization, 9:236--269, 1998. Google ScholarDigital Library
- F. Brglez, X. Y. Li, and M. F. M. Stallmann. On SAT instance classes and a method for reliable performance experiments with SAT solvers. Ann. Math. Artif. Intell., 43(1):1--34, 2005. Google ScholarDigital Library
- M. J. Brusko. Solving personnel tour scheduling problems using the dual all-integer cutting plane. IIE Transactions, 30:835--844, 1998.Google ScholarCross Ref
- Giovanni De Micheli. Synthesis and Optimatization of Digital Circuits. McGraw-Hill, 1994. Google ScholarDigital Library
- D. R. Fulkerson and G. L. Nemhauser. Two computationally difficult set covering problems that arise in computing the 1-width of incidence matrices of Steiner triple systems. Mathematical Programming Study, 2:72--81, 1974.Google ScholarCross Ref
- M. R. Garey and D. S. Johnson. Crossing number is NP-complete. SIAM J. Algebraic and Discrete Methods, pages 312--316, 1983.Google ScholarCross Ref
- Robert S. Garfinkel and George L. Nemhauser. Integer Programming. John Wiley and Sons, 1972.Google Scholar
- Samuel I. Gass. Linear Programming: Methods and Applications. McGraw-Hill, 1958. Google ScholarDigital Library
- R. E. Gomory. Outline of an algorithm for integer solution to linear programs. Bulletin of the American Mathematical Society, 64:275, 1958.Google ScholarCross Ref
- R. E. Gomory. An algorithm for the mixed integer problem. RM-2537. Santa Monica California: Rand Corporation, 1960.Google Scholar
- Gary Hachtel and Fabio Somenzi. Logic Synthesis and Verification Algorithms. Kluwer Academic Publishers, 1996. Google ScholarDigital Library
- ILOG. CPLEX Homepage, 2004. Information on CPLEX is available at http://www.ilog.com/products/cplex/.Google Scholar
- S. Khanna, M. Sudan, L. Trevisan, and D. P. Williamson. The approximability of constraint satisfaction problems. SIAM J. Computing, 30:1863--1920, 2001. Google ScholarDigital Library
- E. L. Lawler, J. K. Lenstra, A. H. G. Rinnooy Kan, and D. B. Shmoys, editors. The Traveling Salesman Problem: A Guided Tour of Combinatorial Optimization. John Wiley, 1985.Google Scholar
- X. Y. Li, M. F. Stallmann, and F. Brglez. Effective Bounding Techniques For Solving Unate and Binate Covering Problems. In Proceedings of the 42nd Design Automation Conference, June 2005. Google ScholarDigital Library
- Xiao Yu Li. Optimization Algorithms for the Minimum-Cost Satisfiability Problem. PhD thesis, Computer Science, North Carolina State University, Raleigh, N. C., August 2004.Google Scholar
- V. M. Manquinho and J. Marques-Silva. On using cutting planes in pseudo-boolean optimization. Journal on Satisfiability, Boolean Modeling and Computation, 2:209--219, 2006.Google ScholarCross Ref
- V. M. Manquinho and J. P. Marques-Silva. Search pruning techniques in SAT-based branch-and-bound algorithms for the binate covering problem. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 21, 2002. Google ScholarDigital Library
- H. M. Salkin and R. D. Koncal. Set covering by an all integer algorithm: Computional experience. JACM, 20:189--193, 1973. Google ScholarDigital Library
- M. Stallmann, F. Brglez, and D. Ghosh. Heuristics, Experimental Subjects, and Treatment Evaluation in Bigraph Crossing Minimization. Journal on Experimental Algorithmics, 6(8), 2001. Google ScholarDigital Library
- M. F. Stallmann and F. Brglez. High-contrast algorithm behavior: Observation, conjecture, and experimental design. Technical Report TR-2007-14, Computer Science Department, NCSU, 2007. Available at http://www.csc.ncsu.edu/research/tech/reports.php.Google Scholar
- Robert J. Vanderbei. Linear Programming: Foundations and Extensions. Kluwer Academic Publishers, 2nd edition, 2001.Google Scholar
- S. Yang. Logic synthesis and optimization benchmarks user guide. Technical Report 1991-IWLS-UG-Saeyang, MCNC, Research Triangle Park, NC, January 1991.Google Scholar
Index Terms
- High-contrast algorithm behavior: observation, hypothesis, and experimental design
Recommendations
Exact Algorithms for the Quadratic Linear Ordering Problem
The quadratic linear ordering problem naturally generalizes various optimization problems such as bipartite crossing minimization or the betweenness problem, which includes linear arrangement. These problems have important applications, e.g., in ...
High-contrast algorithm behavior: observation, conjecture, and experimental design
ecs'07: Experimental computer science on Experimental computer scienceAfter extensive experiments with two algorithms, CPLEX and our implementation of all-integer dual simplex, we observed extreme differences between the two on a set of design automation benchmarks. In many cases one of the two would find an optimal ...
Speeding up IP-based algorithms for constrained quadratic 0---1 optimization
In many practical applications, the task is to optimize a non-linear objective function over the vertices of a well-studied polytope as, e.g., the matching polytope or the travelling salesman polytope (TSP). Prominent examples are the quadratic ...
Comments