M. R. Henzinger, P. Raghavan, and S. Rajagopalan. Computing on data streams. Dimacs Series In Discrete Mathematics And Theoretical Computer Science, pages 107--118, 1999.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....problems require to take decisions without having knowledge, or having only partial knowledge, of future opportunities. Competitive analysis of financial problems has received an increasing attention during the last years, for instance for currency exchange problems [2] or asset al..location [5]. The problem considered in this paper is a generalization of one of the basic on line problems, the ski rental problem due to L. Rudolph (see [4] a model for the well known practical problem rent or buy . The ski rental problem 2 can be stated as follows: you don t know in advance how many ....

P. Raghavan. A Statistical Adversary for On-line Algorithms. DIMACS Series in Discrete Mathematics and Theoretical Computer Science, Vol 7:79-83, 1992.

....set of 1 von Neumann proved his minimax theorem for nite zero sum games. However, since then a number of extensions have been given for continuous games, concave convex games, and games of timing. For a discussion of these and other extensions, the reader is referred to Owen [12] and Raghavan [14]. 7 input sequences for P and let falg 1 ; alg 2 ; g denote the set of deterministic algorithms for P. Then for any any probability distribution x over input sequences and any randomized algorithm algR , R(algR ) min j E x(i) alg j ( i ) opt( i ) # : In many cases, it is ....

P. Raghavan. A statistical adversary for online algorithms. DIMACS Series in Discrete Mathematics and Theoretical Computer Science, 7, 1992.

....problems require to take decisions without having knowledge, or while having only partial knowledge, of future opportunities. Competitive analysis of financial problems has received an increasing attention during the last years, for instance for currency exchange problems [2] or asset al..location [5]. The problem considered in this paper is a generalization of one of the basic online problems, the ski rental problem due to L. Rudolph (see [4] a model for the well known practical problem rent or buy . The ski rental problem can be stated as follows: you don t know in advance how many ....

P. Raghavan. A Statistical Adversary for On-line Algorithms. DIMACS Series in Discrete Mathematics and Theoretical Computer Science, Vol 7:79-83, 1992.

....of problems can gain significantly by including some knowledge of the future. In the areas of finance, economics, and operations research, we find examples of this kind of decision process[8, 16, 21] In this paper, we examine the two way currency trading problem against a statistical adversary[17]. 1.1 Techniques for analyzing on line algorithms The analysis of on line algorithms has typically involved either distributional analysis or competitive analysis. In the former approach, the input is assumed to conform to a natural or typical probability distribution. Based upon this ....

....a powerful adversary often does not reflect the nature of the input to many practical problems. Because the input to most problems lies somewhere between the pessimistic approach of competitive analysis and the more optimistic distributional approach, a number of other approaches were introducted [4, 6, 14, 17, 23]. In this paper we focus on Raghavan s statistical adversary approach[17] Here, the underlying idea is to limit the power of the adversary in some way dependent on the particular problem. Namely, the adversary is required to generate input sequences satisfying certain (statistical) properties. ....

[Article contains additional citation context not shown here]

P. Raghavan. A statistical adversary for on-line algorithms. DIMACS Series in Discrete Mathematics and Theoretical Computer Science, 7:79--83, 1991.

....to do a meaningful worst case analysis of online algorithms without making probabilistic assumptions about the input. The specific financial problems that have been addressed using competitive analysis include foreign currency exchange, mortgage refinancing and portfolio management ( 4] 3] 6] [11]) Recently, rental problems closely related to this work have been studied in [5] and [9] Both papers examine the problem where the purchase cost and the rental cost are fixed. However, it is unknown for what periods of time the item will be needed. In some cases, it may not be worth it to buy ....

....settings, competitive analysis often yields unduly pessimistic results. This is due to the fact that the competitiveness of an algorithm is determined by maximizing over all possible input sequences, and arbitrarily bad inputs do not necessarily occur in practice. In response to this, Raghavan [11] has suggested the use of a statistical adversary who is free to pick any input as long as it obeys certain statistical properties. The statistical adversary is a compromise between probabilistic analysis and competitive analysis where an unrestricted adversary generates the input. The question ....

[Article contains additional citation context not shown here]

P. Raghavan. A Statistical Adversary for On-line Algorithms. DIMACS Series in Discrete Mathematics and Theoretical Computer Science, 7:79-83, 1991.

....fixes an optimal investment portfolio at the start of the game (based on a full knowledge of future events) but it is not allowed to change it thereafter. Note that the on line portfolio selection strategy is very robust; it does not assume anything about the behavior of future events. The paper [Ragh91] analyzes the competitive performance of an on line investment algorithm against a statistical adversary, whose request sequence is required to satisfy certain distributional requirements. Thus, both of these papers restrict the adversary: in the first case, by limiting him to a static policy, ....

P. Raghavan. A statistical adversary for on-line algorithms. On-Line Algorithms, DIMACS Series in Discrete Mathematics and Theoretical Computer Science, 79-83, 1991.

....develop a framework within which these heuristics can be analyzed and propose algorithms that are independent of the input distribution. This framework is based on the notion of competitive analysis. Competitive analysis has been used as a technique to analyze various on line problems [23] 22] [21], 20] 18] 19] In [8] 7] the authors analyze the spin block problem which is similar to the power management problem without any latency consideration. The algorithms and proofs presented in this paper have been adapted from those works. There are signi cant changes in the problem ....

P. Raghavan. A Statistical Adversary for On-line Algorithms. DIMACS Series in Discrete Mathematics and Theoretical Computer Science, 7:79-83, 1991.

No context found.

M. R. Henzinger, P. Raghavan, and S. Rajagopalan. Computing on data streams. Dimacs Series In Discrete Mathematics And Theoretical Computer Science, pages 107--118, 1999.

No context found.

P. Raghavan. A Statistical Adversary for On-line Algorithms. DIMACS Series in Discrete Mathematics and Theoretical Computer Science, Vol 7:79-83, 1991.

No context found.

P. Raghavan. A statistical adversary for on-line algorithms. DIMACS Series in Discrete Mathematics and Theoretical Computer Science, 7:79--83, 1992.

No context found.

P.Raghavan, A statistical adversary for on-line algorithms, DIMACS Series on Discrete Mathematics and Theoretical Computer Science, Vol 7, 1992.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC