Stochastic Dominance and the Bijective Ratio of Online Algorithms - Archive ouverte HAL Access content directly
Journal Articles Algorithmica Year : 2020

Stochastic Dominance and the Bijective Ratio of Online Algorithms


Stochastic dominance is a technique for evaluating the performance of online algorithms that provides an intuitive, yet powerful stochastic order between the compared algorithms. When there is a uniform distribution over the request sequences, this technique reduces to bijective analysis. These methods have been applied in problems such as paging, list update, bin colouring, routing in array mesh networks, and in connection with Bloom filters, and have often provided a clear separation between algorithms whose performance varies significantly in practice. Despite their appealing properties, the above techniques are quite stringent, in that a relation between online algorithms may be either too difficult to establish analytically, or worse, may not even exist. In this paper, we propose remedies to these shortcomings. Our objective is to make all online algorithms amenable to the techniques of stochastic dominance and bijective analysis. First, we establish sufficient conditions that allow us to prove the bijective optimality of a certain class of algorithms for a wide range of problems; we demonstrate this approach in the context of well-studied online problems such as weighted paging, reordering buffer management, and 2-server on the circle. Second, to account for situations in which two algorithms are incomparable or there is no clear optimum, we introduce the bijective ratio as a natural extension of (exact) bijective analysis. Our definition readily generalizes to stochastic dominance. This makes it possible to compare two arbitrary online algorithms for an arbitrary online problem. In addition, the bijective ratio is a generalization of the Max/Max ratio (due to Ben-David and Borodin), and allows for the incorporation of other useful techniques such as amortized analysis. We demonstrate the applicability of the bijective ratio to one of the fundamental online problems, namely the continuous k-server problem on metrics such as the line, the circle, and the star. Among other results, we show that the greedy algorithm attains bijective ratios of O(k) across these metrics.

Dates and versions

hal-02374526 , version 1 (21-11-2019)



Spyros Angelopoulos, Marc Renault, Pascal Schweitzer. Stochastic Dominance and the Bijective Ratio of Online Algorithms. Algorithmica, 2020, 82, pp.1101-1135. ⟨10.1007/s00453-019-00638-w⟩. ⟨hal-02374526⟩
53 View
0 Download



Gmail Facebook X LinkedIn More