A recurring theme in mathematical software evaluation is the generalization of rankings of algorithms on test problems to build knowledge-based recommender systems for algorithm selection. A key issue is to profile al...
详细信息
A recurring theme in mathematical software evaluation is the generalization of rankings of algorithms on test problems to build knowledge-based recommender systems for algorithm selection. A key issue is to profile algorithms in terms of the qualitative characteristics of benchmark problems. In this methodological note, we adapt a novel all-pairs algorithm for the profiling task;given performance rankings for m algorithms on n problem instances, each described with p features, identify a (minimal) subset of p that is useful for assessing the selective superiority of an algorithm over another, for all pairs of m algorithms. We show how techniques presented in the mathematical software literature are inadequate for such profiling purposes. In conclusion, we also address various statistical issues underlying the effective application of this technique.
Although theoretical results for several algorithms in many application domains were presented during the last decades, not all algorithms can be analyzed fully theoretically. Experimentation is necessary. The analysi...
详细信息
Although theoretical results for several algorithms in many application domains were presented during the last decades, not all algorithms can be analyzed fully theoretically. Experimentation is necessary. The analysis of algorithms should follow the same principles and standards of other empirical sciences. This article focuses on stochastic search algorithms, such as evolutionary algorithms or particle swarm optimization. Stochastic search algorithms tackle hard real-world optimization problems, e.g., problems from chemical engineering, airfoil optimization, or bio-informatics, where classical methods from mathematical optimization fail. Nowadays statistical tools that are able to cope with problems like small sample sizes, non-normal distributions, noisy results, etc. are developed for the analysis of algorithms. Although there are adequate tools to discuss the statistical significance of experimental data, statistical significance is not scientifically meaningful per se. It is necessary to bridge the gap between the statistical significance of an experimental result and its scientific meaning. We will propose some ideas on how to accomplish this task based on Mayo's learning model (NPT*).
The field of network science is a highly interdisciplinary area;for the empirical analysis of network data, it draws algorithmic methodologies from several research fields. Hence, research procedures and descriptions ...
详细信息
The field of network science is a highly interdisciplinary area;for the empirical analysis of network data, it draws algorithmic methodologies from several research fields. Hence, research procedures and descriptions of the technical results often differ, sometimes widely. In this paper we focus on methodologies for the experimental part of algorithm engineering for network analysis-an important ingredient for a research area with empirical focus. More precisely, we unify and adapt existing recommendations from different fields and propose universal guidelines-including statistical analyses-for the systematic evaluation of network analysis algorithms. This way, the behavior of newly proposed algorithms can be properly assessed and comparisons to existing solutions become meaningful. Moreover, as the main technical contribution, we provide SimexPal, a highly automated tool to perform and analyze experiments following our guidelines. To illustrate the merits of SimexPal and our guidelines, we apply them in a case study: we design, perform, visualize and evaluate experiments of a recent algorithm for approximating betweenness centrality, an important problem in network analysis. In summary, both our guidelines and SimexPal shall modernize and complement previous efforts in experimental algorithmics;they are not only useful for network analysis, but also in related contexts.
Mining graphs, upon query, for k shortest paths between vertex pairs is a prominent primitive to support several analytics tasks on complex networked datasets. The state-of-the-art method to implement this primitive i...
详细信息
ISBN:
(纸本)9783031785405;9783031785412
Mining graphs, upon query, for k shortest paths between vertex pairs is a prominent primitive to support several analytics tasks on complex networked datasets. The state-of-the-art method to implement this primitive is KPLL, a framework that provides very fast query answering, even for large inputs and volumes of queries, by pre-computing and exploiting an appropriate index of the graph. However, if the graph's topology undergoes changes over time, such index might become obsolete and thus yield incorrect query results. Re-building the index from scratch, upon every modification, induces unsustainable time overheads, incompatible with applications using k shortest paths for analytics purposes. Motivated by this limitation, in this paper, we introduce DECKPLL, the first dynamic algorithm to maintain a KPLL index under decremental modifications. We assess the effectiveness and scalability of our algorithm through extensive experimentation and show it updates KPLL indices orders of magnitude faster than the re-computation from scratch, while preserving its compactness and query performance. We also combine DECKPLL with INCKPLL, the only known dynamic algorithm to maintain a KPLL index under incremental modifications, and hence showcase, on real-world datasets, the first method to support fast extraction of k shortest paths from graphs that evolve by arbitrary topological changes.
This article reports the results of an extensive experimental analysis of efficient algorithms for computing graph spanners in the data streaming model, where an (alpha, beta)-spanner of a graph G is a subgraph S subs...
详细信息
This article reports the results of an extensive experimental analysis of efficient algorithms for computing graph spanners in the data streaming model, where an (alpha, beta)-spanner of a graph G is a subgraph S subset of G such that for each pair of vertices the distance in S is at most a times the distance in G plus beta. To the best of our knowledge, this is the first computational study of graph spanner algorithms in a streaming setting. We compare experimentally the randomized algorithms proposed by Baswana (http://***/abstract?id=oai:***:cs/0611023) and by Elkin (In: Proceedings of the 34th International Colloquium on Automata, Languages and Programming (ICALP 2007), Wroclaw, Poland, pp. 716-727, 9-13 July 2007) for general stretch factors with the deterministic algorithm presented by Ausiello et al. (In: Proceedings of the 15th Annual European Symposium on Algorithms (ESA 2007), Engineering and Applications Track, Eilat, Israel, 8-10 October 2007. LNCS, vol. 4698, pp. 605-617, 2007), designed for building small stretch spanners. All the algorithms we implemented work in a data streaming model where the input graph is given as a stream of edges in arbitrary order, and all of them need a single pass over the data. Differently from the algorithm in Ausiello et al., the algorithms in Baswana (http://***/abstract?id=oai:***:cs/0611023) and Elkin (In: Proceedings of the 34th International Colloquium on Automata, Languages and Programming (ICALP 2007), Wroclaw, Poland, pp. 716-727, 9-13 July 2007) need to know in advance the number of vertices in the graph. The results of our experimental investigation on several input families confirm that all these algorithms are very efficient in practice, finding spanners with stretch and size much smaller than the theoretical bounds and comparable to those obtainable by off-line algorithms. Moreover, our experimental findings confirm that small values of the stretch factor are the case of
In this paper, we study a well-known computationally hard problem, called the subgraph isomorphism problem where the goal is for a given pattern and target graphs to determine whether the pattern is a sub graph of the...
详细信息
In this paper, we study a well-known computationally hard problem, called the subgraph isomorphism problem where the goal is for a given pattern and target graphs to determine whether the pattern is a sub graph of the target graph. Numerous algorithms for solving the problem exist in the literature and most of them are based on the backtracking approach. Since straightforward backtracking is usually slow, many algorithmic refinement techniques are used in practical algorithms. The main goal of this paper is to study such refinement techniques and to determine their ability to speed up backtracking algorithms. To do this we use a methodology of experimental algorithmics. We perform an experimental evaluation of the techniques and their combinations and, hence, demonstrate their usefulness in practice.
This article reports the results of an extensive experimental analysis of efficient algorithms for computing graph spanners in the data streaming model, where an (alpha, beta)-spanner of a graph G is a subgraph S subs...
详细信息
This article reports the results of an extensive experimental analysis of efficient algorithms for computing graph spanners in the data streaming model, where an (alpha, beta)-spanner of a graph G is a subgraph S subset of G such that for each pair of vertices the distance in S is at most a times the distance in G plus beta. To the best of our knowledge, this is the first computational study of graph spanner algorithms in a streaming setting. We compare experimentally the randomized algorithms proposed by Baswana (http://***/abstract?id=oai:***:cs/0611023) and by Elkin (In: Proceedings of the 34th International Colloquium on Automata, Languages and Programming (ICALP 2007), Wroclaw, Poland, pp. 716-727, 9-13 July 2007) for general stretch factors with the deterministic algorithm presented by Ausiello et al. (In: Proceedings of the 15th Annual European Symposium on Algorithms (ESA 2007), Engineering and Applications Track, Eilat, Israel, 8-10 October 2007. LNCS, vol. 4698, pp. 605-617, 2007), designed for building small stretch spanners. All the algorithms we implemented work in a data streaming model where the input graph is given as a stream of edges in arbitrary order, and all of them need a single pass over the data. Differently from the algorithm in Ausiello et al., the algorithms in Baswana (http://***/abstract?id=oai:***:cs/0611023) and Elkin (In: Proceedings of the 34th International Colloquium on Automata, Languages and Programming (ICALP 2007), Wroclaw, Poland, pp. 716-727, 9-13 July 2007) need to know in advance the number of vertices in the graph. The results of our experimental investigation on several input families confirm that all these algorithms are very efficient in practice, finding spanners with stretch and size much smaller than the theoretical bounds and comparable to those obtainable by off-line algorithms. Moreover, our experimental findings confirm that small values of the stretch factor are the case of
It is common in algorithm courses to assess the correctness of algorithms via testing as well as conducting experiments to measure their run-time performance. However, two severe problems must be addressed: both activ...
详细信息
It is common in algorithm courses to assess the correctness of algorithms via testing as well as conducting experiments to measure their run-time performance. However, two severe problems must be addressed: both activities are conducted using different procedures and tools, and algorithm assessment systems do not support well optimization problems due to their unique characteristics. Given the relevant role of optimization algorithms in computing, we present a unified framework to support experimentation with both criteria, that is, optimality and efficiency. The contributions of the article are twofold. First, we present a unified framework to experiment with both criteria, by using explicit principles and by presenting its instantiation in the AlgorEx system. The homogeneous treatment of optimality and time efficiency contributes to smoother integration of experimentation into a course syllabus and to easier adoption of the system. Second, we present our experience in several academic years in an algorithm course. Initially, it was noticed that students had severe difficulties dealing with experimentation, but their academic performance sharply increased by fully integrating experimentation along the complete course. The article also identifies some opportunities for extension in AlgorEx.
GRASP with path-relinking is a hybrid metaheuristic, or stochastic local search (Monte Carlo) method, for combinatorial optimization. A restart strategy in GRASP with path-relinking heuristics is a set of iterations {...
详细信息
GRASP with path-relinking is a hybrid metaheuristic, or stochastic local search (Monte Carlo) method, for combinatorial optimization. A restart strategy in GRASP with path-relinking heuristics is a set of iterations {i (1), i (2), aEuro broken vertical bar} on which the heuristic is restarted from scratch using a new seed for the random number generator. Restart strategies have been shown to speed up stochastic local search algorithms. In this paper, we propose a new restart strategy for GRASP with path-relinking heuristics. We illustrate the speedup obtained with our restart strategy on GRASP with path-relinking heuristics for the maximum cut problem, the maximum weighted satisfiability problem, and the private virtual circuit routing problem.
The efficiency of network virtualization depends on the appropriate assignment of resources. The underlying problem, called virtual network embedding, has been much discussed in the literature, and many algorithms hav...
详细信息
The efficiency of network virtualization depends on the appropriate assignment of resources. The underlying problem, called virtual network embedding, has been much discussed in the literature, and many algorithms have been proposed, attempting to optimize the resource assignment in various respects. Evaluation of those algorithms requires a large number of randomly generated embedding scenarios. This paper presents a novel scenario generation approach and demonstrates how to produce scenarios with a guaranteed exact solution, thereby, facilitating better evaluation of embedding algorithms.
暂无评论