Analyzing alarm floods in a large-scale industrial facility is a sophisticated task, because of too many fault types and their associated consequential alarms. However, there exist similar alarm floods in different pr...
详细信息
Analyzing alarm floods in a large-scale industrial facility is a sophisticated task, because of too many fault types and their associated consequential alarms. However, there exist similar alarm floods in different processes, where the alarms are associated with the same fault types, but configured with different tag names. If these alarm floods are discovered and grouped, the obtained results could facilitate root cause analysis and lead to generalized solutions. Motivated by this practical problem, a systematic pattern-matching method to compare alarm floods across different processes is proposed in this article. The contributions are twofold: 1) A word processing approach is proposed to generalize alarm representations;2) a pattern-matching approach is developed to compare alarm floods across different processes. The effectiveness of the proposed method is demonstrated by a case study using alarm data from an industrial facility.
Scaled matching refers to the problem of finding all locations in the text where the pattern, proportionally enlarged according to an arbitrary real-sized scale, appears. Scaled matching is an important problem that w...
详细信息
Scaled matching refers to the problem of finding all locations in the text where the pattern, proportionally enlarged according to an arbitrary real-sized scale, appears. Scaled matching is an important problem that was originally inspired by Computer Vision. Finding a combinatorial definition that captures the concept of real scaling in discrete images has been a challenge in the patternmatching field. No definition existed that captured the concept of real scaling in discrete images, without assuming an underlying continuous signal, as done in the image processing field. We present a combinatorial definition for real scaled matching that scales images in a pleasing natural manner. We also present efficient algorithms for real scaled matching. The running times of our algorithms are as follows. For T, a two-dimensional nxn text array, and P, an mxm pattern array, we find in T all occurrences of P scaled to any real value in time O(nm (3)+n (2) mlog m).
The reductive decision procedure for unavoidable strings was recently shown to have an exponential lower bound. Hence, as a special case of generalized pattern matching, the existence of an efficient algorithm decidin...
详细信息
The reductive decision procedure for unavoidable strings was recently shown to have an exponential lower bound. Hence, as a special case of generalized pattern matching, the existence of an efficient algorithm deciding string unavoidability remains an interesting open question. It has been hypothesized that some combination of the four necessary conditions implied by the known decidability results would be sufficient. Three of these criteria are determined in polynomial time, and the fourth provides the needed recursion. In this paper, however, we demonstrate the existence of arbitrarily many avoidable strings meeting any extended conjunction of the four necessary conditions. These insufficiency results are achieved by analyzing the appropriate graphical interpretations of the given algorithms. We provide a new combinatorial operation on the corresponding strings and generate arbitrary counterexamples from an empirically located minimal set. Thus, string unavoidability cannot be efficiently decided by the known reductive method or its immediate implications. (c) 2004 Elsevier Inc. All rights reserved.
Let a text string T of n symbols and a pattern string P of m symbols from alphabet Sigma be given. A swapped version T' of T is a length n string derived from T by a series of local swaps (i.e., t(iota)' <-...
详细信息
Let a text string T of n symbols and a pattern string P of m symbols from alphabet Sigma be given. A swapped version T' of T is a length n string derived from T by a series of local swaps (i.e., t(iota)' <-- t(iota+1) and t(iota+1)' <--t(iota)) where each element can participate in no more than one swap. The patternmatching with Swaps problem is that of finding all locations i for which there exists a swapped version T' of T where there is an exact matching of P at location i of T'. It was recently shown that the patternmatching with Swaps problem has a solution in time O(nm(1/3) log m log(2) sigma), where sigma = min(\Sigma\, m). We consider some interesting special cases of patterns, namely, patterns where there is no length-one run, i.e., there are no a, b, c epsilon Sigma where b not equal a and b not equal c and where the substring abe appears in the pattern. We show that for such patterns the patternmatching with Swaps problem can be solved in time O(n log(2) m). (C) 1998 Published by Elsevier Science B.V. All rights reserved.
Scaled matching refers to the problem of finding all locations in the text where the pattern, proportionally enlarged according to an arbitrary integral scale, appears. Scaled matching is an important problem that was...
详细信息
Scaled matching refers to the problem of finding all locations in the text where the pattern, proportionally enlarged according to an arbitrary integral scale, appears. Scaled matching is an important problem that was originally inspired by problems in Vision. However, in real life, a more natural model of scaled matching is the real scaled matching model. Real scaled matching is an extended version of the scaled matching problem allowing arbitrary real-sized scales, approximated by some function, e.g., truncation. It has been shown that the scaled matching problem can be solved in linear time. However, even though there has been follow-up work on the problem, it remained an open question whether real scaled matching could be solved faster than the simple solution of O(nm) time, where n is the text size and m is the pattern size. Using a new approach we show how to solve the real scaled matching problem in linear time. (C) 1999 Elsevier Science B.V. All rights reserved.
暂无评论