This Article 
 Bibliographic References 
 Add to: 
A Novel Discrete Relaxation Architecture
August 1992 (vol. 14 no. 8)
pp. 857-865

The discrete relaxation algorithm (DRA) is a computational technique that enforces arc consistency (AC) in a constraint satisfaction problem (CSP). The original sequential AC-1 algorithm suffers from O(n/sup 3/m/sup 3/) time complexity, and even the optimal sequential AC-4 algorithm is O(n/sup 2/m/sup 2/) for an n-object and m-label DRA problem. Sample problem runs show that these algorithms are all too slow to meet the need for any useful, real-time CSP applications. A parallel DRA5 algorithm that reaches a lower bound of O(nm) (where the number of processors is polynomial in the problem size) is given. A fine-grained, massively parallel hardware computer architecture has been designed for the DRA5 algorithm. For practical problems, many orders of magnitude of efficiency improvement can be reached on such a hardware architecture.

[1] C. G. Bell, "Parallelism advancing computing technologies," Keynote Speech, ACM/IEEE Fall Joint Comput. Conf., Nov. 2-6 1986.
[2] C. G. Bell, "The future of high performance computers in science and engineering,"Comm. ACM, vol. 32, no. 9, pp. 1091-1101, Sept. 1989.
[3] P. R. Cooper and M. J. Swain, "Parallelism and domain dependence in constraint satisfaction," Tech. Rep. TR 255, Dept. of Comput. Sci., Univ. of Rochester, Dec. 1988.
[4] M. J. Flynn, "Some computer organizations and their effectiveness,"IEEE Trans. Comput., vol. C-21, no. 9, pp. 948-960, Sept. 1972.
[5] R. M. Fujimoto, "Simon: A simulator of multicomputer networks," Tech. Rep. ERL Rep., Dept. of EECS, Univ. of California, Berkeley, 1983.
[6] J. Gu, "Parallel algorithms and architectures for very fast search," Tech. Rep. UUCS-TR-88-005, July 1988.
[7] J. Gu, "How to solve Very Large-Scale Satisfiability (VLSS) problems," inJ. Benchmarking SAT Algorithms, Tech. Rep. UCECE-TR-90-002, Oct. 1990.
[8] J. Gu, "A parallel discrete relaxation algorithm and architecture," Tech. Rep. UUCS-TR-88-006, Dept. of Comput. Sci., Univ. of Utah, Mar. 1987 and Jan. 1988; inProc. 3rd IEEE/NASA Symp. Frontiers Massively Parallel Computation(Washington, DC), Oct. 1990, pp. 215-224.
[9] J. Gu and X. Huang, "Implementation and performance of the SAT14 algorithm," to be published.
[10] J. Gu and K. F. Smith, "A structured approach for VLSI circuit design,"IEEE Comput., vol. 22, no. 11, pp. 9-22, Nov. 1989.
[11] J. Gu, W. Wang, and T. C. Henderson, "A parallel architecture for discreie relaxation algorithm," inIEEE Trans. Patt. Anal. Machine Intell., vol. PAMI-9, no. 6, pp. 816-831, Nov. 1987.
[12] R. M. Haralick and L. G. Shapiro, "The consistent labeling problem: Part 1,"IEEE Trans. Patt. Anal. Machine Intell., vol. PAMI-1, no. 2, pp. 173-184, Apr. 1979.
[13] X. Huang and J. Gu, "A quantitative solution for constraint satisfaction," to be published.
[14] K. Hwang, "Advanced parallel processing with supercomputer architectures,"Proc. IEEE, vol. 75, no. 10, pp. 1348-1379, Oct. 1987.
[15] R. R. Johnson, "Critical issues in computer architecture design," Private Communication, 1987-1990.
[16] R. R. Johnson, "Multichip modules: Next-generation packages,"IEEE Spectrum, vol. 27, no. 3, pp. 34-48, Mar. 1990.
[17] W. H. Kautz, M. C. Pease III, and M. W. Green, "Cellular login-in-memory arrays," Tech. Rep. SRI Project 5509, Stanford Res. Inst., Menlo Park, CA, May 1970.
[18] D. C. -L. Ku, "DRA1 chip implementation report," Project Rep., Dept. of Comput. Sci., Univ. of Utah, Mar. 1986.
[19] W. M. Lin and V. K. P. Kumar, "Parallel architectures for discrete relaxation algorithm," inAAAI Workshop on Parallel Algorithms for Artificial Intell.(Detroit), Aug. 20, 1989.
[20] A. K. Mackworth, "Consistency in networks of relations,"Artificial Intell., vol. 8, pp. 99-119, 1977.
[21] J. T. McCall, J. G. Tront, F. G. Gray, R. M. Haralick, and W. M. McCormack, "Parallel computer architectures and problem solving strategies for the consistent labeling problem,"IEEE Trans. Comput., vol. C-34, no. 11, pp. 973-980, Nov. 1985.
[22] C. R. McLean and C. R. Dyer, "An analog relaxation processor," inProc. 5th Int. Conf. Pattern Recogn., 1980, pp. 58-60.
[23] R. Mohr, "A correct path consistency and an optimal generalized arc consistency algorithm," Tech. Rep. CRIN 87-R-030, CRIN-INRIA Lorraine, 1987.
[24] R. Mohr and G. Masini, "Running efficiently are consistency," Tech. Rep, CRIN-INRIA Lorraine, 1986.
[25] R. Mohr and G. Masini, "Good old discrete relaxation," inProc. 8th Euro. Conf. Artificial Intell.(Munchen, Germany), 1988, pp. 651-656.
[26] U. Montanari, "Networks of constraints: Fundamental properties and applications to picture processing,"Inform. Sci., vol. 7, pp. 95-132, 1974.
[27] A. Rosenfeld, R. A. Hummel, and S. W. Zucker, "Scene labeling by relaxation operations,"IEEE Trans. Syst. Man Cybern., vol. SMC-6, no. 6, pp. 420-433, June 1976.
[28] A. Samal, "Parallel split-level relaxation," Dept. of Comput. Sci., Univ. of Utah, Aug. 1988.
[29] C. E. Shannon and W. Weaver,Mathematical Theory of Communication. Urbana, IL: University of Illinois, 1949.
[30] M. J. Swain, "A fast algorithm for arc consistency on a PRAM," Memo, May 22 1988.
[31] M. J. Swain and P. R. Cooper, "Parallel hardware for constraint satisfaction," inProc. Image Understanding Workshop, 1988.
[32] D. Waltz, "Generating semantic descriptions from drawings of scenes with shadows," MAC-AI-TR-271, Mass. Inst. Technol., Cambridge, 1972.
[33] W. Wang, J. Gu, and T. C. Henderson, "A pipelined architecture for parallel image relaxation operations,"IEEE Trans. Circuits Syst., vol. CAS-34, no. 11, pp. 1375-1384, Nov. 1987.

Index Terms:
discrete relaxation algorithm; sequential AC-1 algorithm; time complexity; sequential AC-4 algorithm; parallel DRA5 algorithm; polynomial; computational complexity; parallel algorithms; parallel architectures; polynomials
J. Gu, W. Wang, "A Novel Discrete Relaxation Architecture," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 14, no. 8, pp. 857-865, Aug. 1992, doi:10.1109/34.149596
Usage of this product signifies your acceptance of the Terms of Use.