On the power of randomization in on-line algorithms

@article{BenDavid2005OnTP,
  title={On the power of randomization in on-line algorithms},
  author={Shai Ben-David and Allan Borodin and Richard M. Karp and G{\'a}bor Tardos and Avi Wigderson},
  journal={Algorithmica},
  year={2005},
  volume={11},
  pages={2-14}
}
Against in adaptive adversary, we show that the power of randomization in on-line algorithms is severely limited! We prove the existence of an efficient “simulation” of randomized on-line algorithms by deterministic ones, which is best possible in general. The proof of the upper bound is existential. We deal with the issue of computing the efficient deterministic algorithm, and show that this is possible in very general cases. 

On the Power of Randomization in Distributed Algorithms in Dynamic Networks with Adaptive Adversaries

It is proved that randomness offers limited power to better deal with “bad” adaptive adversary, and that P can always be converted to a new algorithm Q with comparable time complexity, even when Q runs against prophetic adversaries.

A 4/3-competitive randomized algorithm for online scheduling of packets with agreeable deadlines

This work claims that the phi-competitive deterministic online algorithm by Li et al. can be slightly simplified, while retaining its competitive ratio, and argues that the competitive ratio against oblivious adversary is at most 4/3.

A 4/3-competitive randomized algorithm for online scheduling of packets with agreeable deadlines

This work claims that the algorithm by Li et al. can be slightly simplified, while retaining its competitive ratio, and argues that the competitive ratio against oblivious adversary is at most 43.

A Universal Randomized Packet Scheduling Algorithm

ReMix is the optimal memoryless randomized algorithm against adaptive adversary for Packet Scheduling, and its general analysis yields improved performance guarantees for several restricted variants, including the s-bounded instances.

Randomized algorithm for the k-server problem on decomposable spaces

A method is given which extends existing algorithms to larger spaces with the growth rate of the competitive quotients being at most O (log k) and yields o ( k )-competitive algorithms solving the randomized k -server problem for some special underlying metric spaces, e.g. HSTs of “small” height (but unbounded degree).

Randomized algorithm for the k-server problem on decomposable spaces

Randomized Priority Algorithms: (Extended Abstract)

This work extends the model of Borodin, Nielsen and Rackoff so as to formally define "randomized greedy-like algorithms" and be able to prove lower bounds on the proximability of a certain problem by such a class of algorithms.

Memoryless Algorithms for the Generalized k-server Problem on Uniform Metrics

It is shown that the Harmonic Algorithm achieves this competitive ratio and provides matching lower bounds, which improves the doubly-exponential bound of Chiplunkar and Vishwanathan for the more general setting of uniform metrics with different weights.

One to Rule Them All: A General Randomized Algorithm for Buffer Management with Bounded Delay

We give a memoryless scale-invariant randomized algorithm Mix-R for buffer management with bounded delay that is e/(e - 1)- competitive against an adaptive adversary, together with better performance
...

References

SHOWING 1-10 OF 17 REFERENCES

A strongly competitive randomized paging algorithm

The partitioning algorithm is developed, a randomized on-line algorithm for the paging problem, which it is proved that its expected cost on any sequence of requests is within a factor ofHk of optimum.

Memory Versus Randomization in On-line Algorithms (Extended Abstract)

The amortized analysis of LRU shows that it achieves optimal worst-case performance (this will be made precise in section 2).

Infinite games: randomization, computability, and applications to online problems

It is shown that, whereas all open games are semicompu tably determinate, there is a semicomputab]y indeterminate e closed game, and the proof for the indeterminacy result requires a new diagonalization technique, which might be useful in other similar cases.

A competitive 3-server algorithm

This work proves that a simple, natural randomized algorithm is competitive for the 3-server problem, where to serve a request is to move one of the k servers to the request site.

Competitive k-server algorithms

Deterministic competitive k-server algorithms are given for all k and all metric spaces and the competitive ratio can be proved is exponential in the number of servers, settling the k- server conjecture.

The harmonic online K-server algorithm is competitive

The Harmonic algorithm for the online K-server problem is shown to be competitive against an adaptive online adversary for K 22 and the best competitive ratios that have been published so far for online algorithms over a general metric space when K >2 are shown.

New results on server problems

A fast algorithm for oflline computing of an optimal schedule is given, and it is shown that finding an optimal offline schedule is at least as hard as the assignment problem.

Competitive algorithms for on-line problems

This paper presents several general results concerning competitive algorithms, as well as results on specific on-line problems.

The harmonic k-server algorithm is competitive

A simple proof that the Harmonic Harmonic k-server algorithm is competitive is given, and the competitive ratio is proved to be the best currently known fo the algorithm.