ABSTRACT
A critical computational requirement for many of the decision technologies in the fields of operations research (PERT/CPM, Markov chains, decision trees, Bayesian analysis, MRP, simulation, …), artificial intelligence (evidential reasoning, truth maintenance systems, propositional logic, rule based inference, frames and semantic nets, …), and decision support systems (worksheet or financial planning models, data / entity models, …) is the development and manipulation of a function network or directed graph describing the relationship between "variables", "objects", or "actors" involved in the application of the decision technology to a specific problem. The manipulation of such networks or graphs using Boolean matrices and vector of integer vectors is well known in portions of the APL community (see bibliography), but intertwined with specific applications and spread out across a variety sources (some of which are difficult to obtain). This paper succinctly and simply describes the basics of manipulating a function network with Boolean matrices and integer vectors including focusing networks, finding circular conditions (A depends on B, B depends on C, C depends on A, therefore A depends on A, …), and grouping functions based on relative independence to identify parallel computational opportunities and substantially reduces the non-procedural aspect of the problem.
- 1.Alfonseca, M. and Brown, J., "Parallel Solutions to Logic Problems" SEAS Spring Meeting 1987, SEAS Proceedings, Vol. I, p. 27-46; and reprinted in APL-CAM Journal, Vol. 9:4, p. 764-778, (1987).Google Scholar
- 2.Brown, J., Pakin, S., and Polivka, R., APL2 at a Glance, Prentice Hall, Englewood New Jersey (1988). Google ScholarDigital Library
- 3.Eusebi, E., "Inductive Reasoning from Relations", APL87 Conference Proceedings, APL Quote Quad, Vol. 17, No. 4, pp. 386-390 (1987). Google ScholarDigital Library
- 4.Fordyce, K. Morreale, M., McGrew, J. and Sullivan, G., "APL Techniques in Knowledge Based Systems", IBM, 33VA/284, Kingston, NY 12401 forthcoming in Encyclopedia of Computer Science and Technology, edited by Allen Kent and James William from the University of Pittsburgh (1991).Google Scholar
- 5.Fordyce, K., Jantzen, J., and Sullivan, G. Sr.,and Sullivan, G. Jr., "Representing Knowledge with Functions and Boolean Arrays" IBM Journal of Research and Development Vol. 33, No. 6, pp. 627-646 (1989). Google ScholarDigital Library
- 6.Fordyce, K., and Sullivan, G. "Boolean Array Based Inference Engines", Proceedings of the APL and Expert Systems Conference, Syracuse, NY, a copy may be obtained from the authors (1988).Google Scholar
- 7.Fordyce, K. and Sullivan, G., "Boolean Array Structures for a Rule Based Forward Chaining Inference Engine", APL Quote Quad, APL87 Conference Proceedings, Vol. 17, No. 4, pp. 185-195 (1987). Google ScholarDigital Library
- 8.Franksen, O., Falster, P., and Evans, F., "Qualitative Aspects of Large Scale Systems - Developing Design Rules Using APL", Lecture Notes in Control and Information Sciences (Vol. 17), Monograph, Spfinger-Verlag (1979). Google ScholarDigital Library
- 9.Franksen, O., "Group Representation of Finite Polyvalent Logic- A case Study Using APL Notation", in A Link Between Science and Application of Automatic Control, Proceedings IFAC World Congress 1978, edited by A. Niemi, Pergamon Press, New York, Vol 2., pp. 875-887. Monograph, Springer-Verlag (1979).Google Scholar
- 10.Franksen, O., "Are Data Structures Geometrical Objects? Part 1: Invoking the Erlanger Program", Syst. Anal. Model. Simul., Vol. 1, No. 2, pp. 113-130 (1984).Google Scholar
- 11.Franksen, O., "Are Data Structures Geometrical Objects? Part 2: Invariant Forms in APL and Beyond", Syst. Anal. Model. Simul., Vol. 1, No. 2, pp. 131-150 (I 984).Google Scholar
- 12.Franksen, O., "Are Data Structures Geometrical Objects? Part 3: Appendix A: Linear Differential Operators", Syst. Anal. Model. Simul., Vol. 1, No. 3, pp. 249-258 (1984).Google Scholar
- 13.Franksen, O., "Are Data Structures Geometrical Objects? Part 4: Appendix B" Logic Invariants by Finite Truthtables", Syst. Anal. Model. Simul., Vol. 1, No. 4, pp. 339-350 (1984).Google Scholar
- 14.Franksen, O., Mr. Babbage's Secret. The Tale of a Cypher- and APL, Prentice Hall, New Jersey (1985). Google ScholarDigital Library
- 15.lverson, K., "1979 ACM Turing Award Lecture: Notation as a Tool for Thought", Communications of the ACM, Vol. 23, No. 8, pp. 444-465 (1980). Google ScholarDigital Library
- 16.Jantzen, J., "Inference Planning Using Digraphs and Boolean Arrays", APL Quote Quad, APL89 Conference Proceedings, Vol. 19, No. 4. pp. 200-204 (1989). Google ScholarDigital Library
- 17.Metzger, R., "Using Graphs to Analyze APL Functions", APL Quote Quad (APL83 Conference Proceedings), Vol. 13, No. 3. pp. 153-161 (1983). Google ScholarDigital Library
- 18.Moiler, G., "A Logic Programming Tool for Qualitative System Design", APL Quote Quad, APL86 Conference Proceedings, Vol. 16, No. 4, pp. 266-71 (1986). Google ScholarDigital Library
- 19.O'rithm, A., "Directed Graphs", NY SIG APL Newsletter, January 1991, February 1991 (pp. 4-6), March 1991 (pp. 2-4) ( 1991).Google Scholar
- 20.Tarjan, R., "Testing Flow Graph Reducibility", Journal of Computer and Systems Sciences, Vol. 9, No. 3, (December 1974).Google ScholarDigital Library
- 21.Thomson, N., APL Programs for the Mathematical Classroom, Springer Verlag, New York, IS BN 0-387-97002-9 (1989). Google ScholarDigital Library
- 22.Thomson, N., "APL2 and Basic Operation Research Algorithms", Proceedings of Share European Association (SEAS) Fall 1989 Meeting, Vol. 2, pp. 1689-1703 IBM UK Labs, Mail Point 188, Hursley House, Hursley Park, Winchester, Hampshire S021 21 N, England (1989).Google Scholar
Index Terms
- Using boolean of integer arrays to analyze networks
Recommendations
Using boolean of integer arrays to analyze networks
A critical computational requirement for many of the decision technologies in the fields of operations research (PERT/CPM, Markov chains, decision trees, Bayesian analysis, MRP, simulation, …), artificial intelligence (evidential reasoning, truth ...
Zero/Positive Capacities of Two-Dimensional Runlength-Constrained Arrays
A binary sequence satisfies a one-dimensional $(d_1, k_1, d_2, k_2)$ runlength constraint if every run of zeros has length at least $d_1$ and at most $k_1$ and every run of ones has length at least $d_2$ and at most $k_2$ . A two-dimensional binary ...
Using simulation to analyze supply chains
WSC '00: Proceedings of the 32nd conference on Winter simulationSupply Chain management, the management of the flow of goods or services from materials stage to the end user, is a complex process because of the level of uncertainty at each stage of the supply chain. Computer simulation, because it can be applied to ...
Comments