Elias Koutsoupias and Christos H. Papadimitriou. Beyond competitive analysis. SIAM Journal on Computing, 30(1):300--317, 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....various extensions and alternatives to pure competitive analysis have been investigated in the literature. In comparative analysis the class of algorithms where the offline algorithm comparative is chosen from is restricted. This concept has been introduced in the context of the paging problem [KP94] The authors compare the performance of an online algorithm for the paging problem with that of the best paging algorithm having limited lookahead. Let # be a minimization (online) problem. The comparative ratio c ALG,B of an algorithm ALG for # relative to a class comparative ratio of ....

....an online algorithm is the concept of resource augmentation (see e.g. PS 97, PK95, ABF96, ST85] Here, resource the online algorithm is given more resources (e.g. more or faster machines in scheduling) to serve requests than the offline adversary. The diffuse adversary model introduced in [KP94] deals with the situation diffuse adversary where the input is chosen by an adversary according to some probability distribution. Although the online algorithm does not know the distribution itself, it is given the information that this distribution belongs to a specific class of distributions. ....

E. Koutsoupias and C. Papadimitriou, Beyond competitive analysis, Proceedings of the 35th Annual IEEE Symposium on the Foundations of Computer Science, 1994, pp. 394--400.

....algorithms with partial information, which was introduced by Papadimitriou and Yannakakis [54] in the context of linear programming; their model corresponds to a distributed system with no communication. A generalization of this approach has recently been described by Koutsoupias and Papadimitriou [43]. In addition, there is a long history of interest in optimality of a distributed algorithm given certain conditions, such as a particular pattern of failures [27, 31, 37, 50, 52, 53] or a particular pattern of message delivery [12, 33, 55] These and related works are in the spirit of our ....

E. Koutsoupias and C. Papadimitriou. Beyond Competitive Analysis. In Proc. 33rd IEEE Symposium on Foundations of Computer Science, pp. 394-400, November 1994.

....appealling [7] A method to weaken worst case analysis is to impose additional constraints on the adversary. The idea of constrained adversaries is well known in Mathematical Economics [8, 10] and Control Theory [2] and it has been recently and independently introduced into Computer Science [13]. The main idea behind constrained adversaries is that the input instance is generated according to a worst case distribution chosen from a given class of distributions. In other words, the adversary has to satisfy an additional set of constraints that define the set of feasible input ....

E. Koutsoupias and C. H. Papadimitriou. Beyond competitive analysis. SIAM Journal on Computing, 30(1):394-400, 2000.

....will mention related work relevant to the specific problems we consider. Here, we o#er a brief overview of generally relevant related work. The importance of when information becomes available has been noted previously, especially in the significant body of work on algorithms with lookahead, e.g. [12, 22, 24, 28]. In the case of on line decision models, however, the possibility of not having up to date information is not generally addressed. For load balancing problems, the question has been considered for statistical models [30, 31, 37] And recently, 5] considers an on line load balancing setting where ....

E. Koutsoupias and C.H. Papadimitriou. Beyond competitive analysis. In In Proc. 35th Ann. Symp. on Foundations of Computer Science, pp. 394--400, 1994.

....For example, the access graph model [5, 14, 10] restricts the possible choices of the next request as the function of the current request, so as to model locality of reference. In contrast, some researchers have tried to enhance the power of the on line algorithm, for example, with lookaheads [24, 16]. Other researchers have suggested alternative evaluation criteria. For exam ple, Ben David and Borodin [3] indicate that some competitive algorithms re quire unbounded memory, and that finite lookahead is useless for improving the competitive ratio. They suggested the max max ratio as an ....

Elias Koutsoupias and Christos H. Papadimitriou. Beyond competitive analysis. In 35th Annual Symposium on Foundations of Computer Science, pages 394-400, Santa Fe, New Mexico, 20-22 November 1994. IEEE.

....to the model used by Gafni and Mitzenmacher [23] in their analysis of mutual exclusion protocols with random timing, but is extended to include constant delays inserted by the adversary in addition to random delays. Another source of inspiration is Koutsoupias and Papadimitriou s di use adversary [24], which chooses a distribution over executions in which no branch at any decision point can occur with probability more than some xed . Our model is not the rst in which an adversary chooses parameters for a stochastic process that then controls scheduling; a sophisticated model of this type, ....

Elias Koutsoupias and Christos H. Papadimitriou. Beyond competitive analysis. In 35th Annual Symposium on Foundations of Computer Science, pages 394-400, Santa Fe, New Mexico, 20-22 November 1994. IEEE.

....to the model used by Gafni and Mitzenmacher [23] in their analysis of mutual exclusion protocols with random timing, but is extended to include constant delays inserted by the adversary in addition to random delays. Another source of inspiration is Koutsoupias and Papadimitriou s di use adversary [24], which chooses a distribution over executions in which no branch at any decision point can occur with probability more than some xed . Our model is not the rst in which an adversary chooses parameters for a stochastic process that then controls scheduling; a sophisticated model of this type, ....

Elias Koutsoupias and Christos H. Papadimitriou. Beyond competitive analysis. In 35th Annual Symposium on Foundations of Computer Science, pages 394-400, Santa Fe, New Mexico, 20-22 November 1994. IEEE.

....9 Acknowledgments Some of the material in this survey has been adapted from [AADW94] AW96] and [AH96] None of this work would have been possible without the efforts of my co authors Miklos Ajtai, Cynthia Dwork, Will Hurwood, and Orli Waarts. 7 The technique of comparative analysis [KP94] might seem to be the right answer to these problems. Comparative analysis measures the value of information by comparing the best algorithms in two information regimes (e.g. paging algorithms with l lookahead versus paging algorithms with no lookahead) A complication, from the point of view ....

E. Koutsoupias and C. Papadimitriou. Beyond competitive analysis. In Proc. 25th Symposium on Foundations of Computer Science, pages 394-- 400, 1994.

....situation various extensions and alternatives to pure competitive analysis have been investigated in the literature. In comparative analysis the class of algorithms where the offline algorithm is chosen from is restricted. This concept has been explored in the context of the Paging Problem [34] and the Online TSP [14] Another approach to strengthen the position of an online algorithm is the concept of resource augmentation (see e.g. 41, 42, 10, 46] Here, the online algorithm is given more resources (more or faster machines in scheduling) to serve requests than the offline adversary. ....

....Another approach to strengthen the position of an online algorithm is the concept of resource augmentation (see e.g. 41, 42, 10, 46] Here, the online algorithm is given more resources (more or faster machines in scheduling) to serve requests than the offline adversary. The diffuse adversary [34] model deals with the situation where the input is chosen by an adversary according to some probability distribution. Although the online algorithm does not know the distribution itself, it is given the information that this distribution belongs to a specific class of distributions. Other ....

E. Koutsoupias and C. Papadimitriou, Beyond competitive analysis, Proceedings of the 35th Annual IEEE Symposium on the Foundations of Computer Science, 1994, pp. 394--400.

....against a fair adversary. Moreover, no deterministic online algorithm can achieve a competitive ratio smaller than 1 p 17 4 against the fair adversary. The use of a restricted adversary falls within the concept of comparative analysis, which was introduced by Koutsoupias and Papadimitriou [14]. The authors compare the performance of an online algorithm for the Paging Problem with that of the best paging algorithm having limited lookahead. Let P be a minimization (online) problem. The comparative ratio of an algorithm ALG for P relative to a class B of algorithms is defined as the worst ....

E. Koutsoupias and C. Papadimitriou, Beyond competitive analysis, Proceedings of the 35th Annual IEEE Symposium on the Foundations of Computer Science, 1994, pp. 394--400.

....Inputs Neal E. Young Abstract In evaluating an algorithm, worst case analysis can be overly pessimistic. Average case analysis can be overly optimistic. An intermediate approach is to show that an algorithm does well on a broad class of input distributions. Koutsoupias and Papadimitriou [10] recently analyzed the least recently used (Lru) paging strategy in this manner, analyzing its performance on an input sequence generated by a so called di use adversary one that must choose each request probabilitistically so that no page is chosen with probability more than some xed ....

.... considered (e.g. 3, 4, 8, 6, 9, 11, 16, 17] For a survey on competitive analysis of paging, we refer the reader to the recent book by Borodin and El Yaniv [2, ch s 3 5] This paper concerns the following generalization of the standard model, recently proposed by Koutsoupias and Papadimitriou [10]. For any class of distributions on the input sequences and any deterministic or randomized algorithm A, de ne R( A) the competitive ratio of A against the di use adversary, to be the minimum (again, to be precise, in mum) c such that for each distribution D in , there is a constant b such ....

[Article contains additional citation context not shown here]

Elias Koutsoupias and Christos H. Papadimitriou. Beyond competitive analysis. In Proc. of the 35th IEEE Annual Symp. on Foundation of Computer Science, pages 394-400, 1994.

No context found.

Elias Koutsoupias and Christos Papadimitriou. Beyond competitive analysis. SIAM Journal on Computing, 30(1):300-317, 2000.

....although there are other algorithms that may also be optimal. An important byproduct of our analysis is that, extending the work in [9] we completely characterize the work functions of the paging special case of the k server problem. In a preliminary version of this work that appeared in [8], we incorrectly stated that the competitive ratio R( is given by a simple Markov chain of (k 1 ) k 1 states. In fact, the answer is more complicated. The competitive ratio is given by the optimal Markov chain from a family of k (k 1 ) k 1 Markov chains of (k 1 ) k ....

....next p n numbers has probability at least 1= p n to end the phase. Elaborating on this observation we get that R is in the interval [1 p =2; 1 2 p ] Numerical evaluations suggest that the value of R is approximately 1 0:8 p , when 0. A preliminary version of this work [8] had an incorrect proof of optimality of LRU. The proof was based on the unjusti ed assumption that a conservative adversary that achieves optimal competitive ratio against LRU assigns probability to exactly one page in the on line fast memory. As was pointed out to us by Neal Young (see also ....

E. Koutsoupias and C. H. Papadimitriou, Beyond competitive analysis, in Proceedings 35th Annual Symposium on Foundations of Computer Science, 1994, pp. 394-400.

No context found.

Elias Koutsoupias and Christos H. Papadimitriou. Beyond competitive analysis. SIAM Journal on Computing, 30(1):300--317, 2000.

No context found.

Elias Koutsoupias and Christos H. Papadimitriou. Beyond competitive analysis. SIAM Journal on Computing, 30(1):300--317, 2000.

No context found.

E. Koutsoupias and C. Papadimitriou, \Beyond competitive analysis", SIAM Journal on Computing 30, pp. 300-317, 2000.

No context found.

E. Koutsoupias and C. H. Papadimitriou. Beyond competitive analysis. SIAM Journal on Computing, 30(1):300-- 317, 2000.

No context found.

E. Koutsoupias and C. H. Papadimitriou, "Beyond Competitive Analysis", SIAM Journal on Computing, Vol. 30, No. 1, pp. 300--317, April/May 2000.

No context found.

E. Koutsoupias and C. H. Papadimitriou, "Beyond Competitive Analysis", SIAM Journal on Computing, Vol. 30, No. 1, pp. 300--317, April/May 2000.

No context found.

Koutsoupias E and Papadimitriou C. Beyond competitive analysis. SIAM Journal on Computing 2000; 30:300-317.

No context found.

E. Koutsoupias and C. Papadimitriou. Beyond competitive analysis. In Proceedings of the 35th Annual IEEE Foundations of Computer Science, pages 394-400, 1994.

No context found.

E. Koutsoupias and C. Papadimitriou. Beyond competitive analysis. In Proceedings of the 35th Annual IEEE Foundations of Computer Science, pages 394-400, 1994.

No context found.

E. Koutsoupias and C. H. Papadimitriou. Beyond competitive analysis. In 35th Annual Symposium on Foundations of Computer Science, pages 394--400, 1994.

No context found.

E. Koutsoupias and C. Papadimitriou. Beyond competitive analysis. In Proceedings of the Twenty-Fifth Symposium on Foundations of Computer Science, pages 394{ 400, 1994.

No context found.

E. Koutsoupias and C. Papadimitriou. Beyond competitive analysis. In Proceedings 1994.

First 50 documents  Next 50

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