Showing papers on "Local search (optimization) published in 1979"
TL;DR: A library containing the spectra of nearly 17 000 organic compounds is used with these methods to provide an evaluation of the relative performance of different search techniques.
Abstract: Masa spectral search algorithms are tested by two methods. Trial searches of target spectra are used to compare the effects of various data-encoding methods and different distance metrics on search results. The effectiveness of selected search algorithms in retrieving similar compounds is assessed through the use of a propagation method, which employs repetitive searches of the nearest matches to a “seed” compound. A library containing the spectra of nearly 17 000 organic compounds is used with these methods to provide an evaluation of the relative performance of different search techniques.
66 citations
TL;DR: A local search algorithm for the design of small-diameter networks is presented for both directed and undirected regular graphs, and the resulting graphs are at least as good as any previously known.
Abstract: A local search algorithm for the design of small-diameter networks is presented for both directed and undirected regular graphs. In all cases the resulting graphs are at least as good as any previously known, in the sense that they have at least as small a diameter and average shortest distance for a given number of nodes and degrees.
30 citations
01 Jan 1979
TL;DR: The best-bound-first search strategy is shown to be optimal in both time and space and the probabilistic model of its domain is given, namely a class of trees with an associated probability measure.
Abstract: : Many important problems in operations research, artificial intelligence, and other areas of computer science seem to require search in order to find an optimal solution. A branch and bound procedure, which imposes a tree structure on the search, is often the most efficient known means for solving these problems. While for some branch and bound algorithm a worst case complexity bound is known, the average case complexity is usually unknown despite the fact that it gives more information about the performance of the algorithm. In this dissertation the branch and bound method is discussed and a probabilistic model of its domain is given, namely a class of trees with an associated probability measure. The best-bound-first search strategy and depth-first search strategy are discussed and results on the expected time and space complexity of these strategies are presented and discussed. The best-bound-first search strategy is shown to be optimal in both time and space. These results are illustrated by data from randomly generated traveling salesman problems.
12 citations
30 May 1979
TL;DR: This paper describes a general search procedure for finding a best member of a large discrete set of solutions of APL functions that dynamically expands and contracts the scope of the search.
Abstract: This paper describes a general search procedure for finding a best member of a large discrete set of solutions. The Traveling Salesperson Problem illustrates how the method is used. Two APL functions contain the core of the method. NEXTSET implements the user's definition of the neighborhood of a solution while WHATNOW expresses the search strategy to be employed. WHATNOW can incorporate any calculations, or system or session parameters available to the function as the search progresses, and it passes its determination on to NEXTSET which dynamically expands and contracts the scope of the search.