Showing papers on "Breadth-first search published in 1990"

Buffer placement in distributed RC-tree networks for minimal Elmore delay

Abstract: An algorithm is presented for choosing the buffer positions for a wiring tree such that the Elmore delay is minimal. For given required arrival times at the sinks of the wiring tree, the algorithm chooses buffers such that the required departure time at the source is as late as possible. The topology of the wiring tree, a Steiner tree, is assumed to be given, as well as the possible (legal) positions of the buffers. The algorithm uses a depth first search on the wiring tree to construct a set of time/capacitance pairs that correspond to different choices. The complexity of the algorithm is O(B/sup 2/), where B is the number of possible buffer positions. An extension of the basic algorithm allows minimization of the number of buffers as a secondary objective. >

Adjacency constraints in harvest scheduling: an aggregation heuristic

Abstract: An heuristic for adjacency constraint aggregation is proposed. The heuristic is composed of two procedures. Procedure 1 consists of identifying harvesting areas for which it is not necessary to wri...

Parallel graph algorithms based upon broadcast communications

Abstract: Some common guidelines that can be used to design parallel algorithms under the single-channel broadcast communication model are presented. Several graph problems are solved, including topological ordering, the connected component problem, breadth-first search, and depth-first search. If an ideal conflict resolution scheme is used, all of the algorithms require O(n) time by using n processors. Under such a situation, the algorithms are all optimal. If a realistic conflict resolution is used, the algorithms require O(n log n) time by using n/log n processors. For both cases, all of the algorithms achieve optimal speedups. >

Thermal generator scheduling algorithm based on heuristic-guided depth-first search

Abstract: This paper develops an algorithm which is suitable for implementation on computers by a logic programming language, Prolog, for scheduling thermal generators to meet the daily rising system demand and spinning reserve requirement in the operation of power systems. The algorithm is based on the depth-first and heuristic search techniques in artificial intelligence, and an interpretation of the scheduling process as a tree searching problem. Heuristic rules are derived for overcoming the problem of combinatorial explosion in the search space of the scheduling process and for quickly determining the solution schedule having the lowest total excessive generation capacity between the total generation capacity and the combined system demand and spinning reserve. The ramp rate characteristics and crew constraints of generator units together with the system constraint of must-on units are incorporated in the algorithm. The application of the developed algorithm which is implemented by Prolog to schedule ten thermal generators is presented.

Parallel algorithms for geometric intersection graphs

Abstract: This dissertation presents efficient parallel algorithms on the PRAM for solving several problems on the interval graphs, circular-arc graphs, and intersection graphs of rectangles, assuming that input graphs are represented by intersection models, instead of standard graph representations such as adjacency matrices and edge lists. Specifically, our results are the following: For interval graphs, we present O(log n) time, n/log n processor algorithms in the EREW PRAM for the problems of computing a depth first search tree, a breadth first search tree, a maximum independent set, a minimum clique cover, a minimum dominating set, and a minimum coloring, assuming that the intervals are initially sorted by x-coordinates. For circular-arc graphs, we give O(log n) time, n/log n processor algorithms in the EREW PRAM for the problems of finding a depth first search tree, a breadth first search tree, a minimum circle cover, a maximum independent set, and a minimum dominating set, assuming the circular-arcs are initially sorted by angles. An O(log$\sp2$n) time, O($n\sp3$/log n) processor algorithm in the CREW PRAM is given for finding a maximum clique. For a set of rectangles, we present O(log n) time, n processor algorithms in the CREW PRAM for the problems of finding a maximum clique and the problems of computing the area and perimeter of the union of rectangles. We also present an O(log n) time, n processor algorithm in the CRCW PRAM for the problem of computing the connected components of the rectangles. Finally, we discuss three geometric problems, namely, region labeling of binary images, finding largest rectangles in a black-and-white grid, and segment-dragging problem, and present efficient parallel solutions for them.

An algorithm for the classification of strongly regular graphs by means of lexically ordered adjacency matrices

Abstract: The concept of lexically ordered adjacency matrix of a graph is introduced and it is proved that every adjacency matrix is isomorphic to at least one lexically ordered adjacency matrix. An algorithm for the classification of strongly regular graphs is developed, where the property of lexical ordering is used as a means to reduce the number of generated adjacency matrices. We also describe other pruning methods that can be used.