Uniformly distributed pseudo-random number generators (pseudo-RNGs) are at the base of numerous problems (such as those in operations research and machine learning) whose deterministic solution is often exponential/co...
详细信息
Uniformly distributed pseudo-random number generators (pseudo-RNGs) are at the base of numerous problems (such as those in operations research and machine learning) whose deterministic solution is often exponential/combinatorial and hence intractable. While designing perfect RNGs is impossible, it is possible to select one among the existing RNGs, which would be the core of a probabilistic algorithm to solve such a problem in polynomial-time and produce the best quality result. Presented here is a study of only three RNGs and their merits just to highlight an indirect way of comparing any two or more RNGs before selecting one to be used in the probabilistic algorithm. The indirect approach depends on the performance of an RNG in the specified randomized algorithm used for test problems. How good the quality of results is and how less the computational complexity is, are the core of this approach. A direct approach for comparing two or more RNGs, on the other hand, is independent of algorithms. We use in this approach one or more testing procedures for random numbers to decide which one we should choose. We believe that such a study would help one in choosing an appropriate RNG for his problem. (C) 2004 Elsevier Ltd. All rights reserved.
This paper focuses on protecting harbor by monitoring the behavior of ships moving in the harbor. In particular, we propose to combine a probabilistic Hidden Markov Model approach with a formal language approach to mo...
详细信息
ISBN:
(纸本)9781467308298
This paper focuses on protecting harbor by monitoring the behavior of ships moving in the harbor. In particular, we propose to combine a probabilistic Hidden Markov Model approach with a formal language approach to model and recognize usual and unusual ship behavior. The probabilistic approach was mainly used to recognize usual ship behavior. However, the reactive approach was used to recognize abnormal ship behavior and classify it into aggressive and transgressive. By combining probabilistic and reactive approaches, we conceived a Behavior Analysis Modulus that we called BAM. It is a decision support system for port team security that aims to generate security advice on ships as they enter into the harbor.
A greedy randomized adaptive search procedure (GRASP) is a metaheuristic for combinatorial optimization. In this paper, we describea GRASP for the graph planarization problem, extending the heuristic of Goldschmidt an...
详细信息
暂无评论