Joaquín Pérez

Automatic Generation of Control Parameters for the Threshold Accepting Algorithm

In this article a new method to obtain the control parameters values for the Threshold Accepting algorithm is presented, which is independent of the problem domain and size. This approach differs from the traditional methods that require knowing first the problem domain, and then knowing how to select the parameters values to solve specific problem instances. The proposed method is based on a sample of problem instances, whose solution allows us to characterize the problem and to define the parameters. To test the method the combinatorial optimization model called DFAR was solved using the Threshold Accepting algorithm. The experimental results show that it is feasible to automatically obtain the parameters for a heuristic algorithm, which will produce satisfactory results, even though the kind of problem to solve is not known. We consider that the proposed method principles can be applied to the definition of control parameters for other heuristic algorithms.

Vertical Fragmentation and Allocation in Distributed Databases with Site Capacity Restrictions Using the Threshold Accepting Algorithm

This paper presents an extension of the DFAR mathematical optimization model, which unifies the fragmentation, allocation and dynamical migration of data in distributed database systems. The extension consists of the addition of a constraint that models the storage capacity of network sites. This aspect is particularly important in large databases, which exceed the capacity of one or more sites. The Threshold Accepting Algorithm is a variation of the heuristic method known as Simulated Annealing, and it is used for solving the DFAR model. The paper includes experimental results obtained for large test cases.

A Statistical Approach for Algorithm Selection

This paper deals with heuristic algorithm characterization, which is applied to the solution of an NP-hard problem, in order to select the best algorithm for solving a given problem instance. The traditional approach for selecting algorithms compares their performance using an instance set, and concludes that one outperforms the other. Another common approach consists of developing mathematical models to relate performance to problem size. Recent approaches try to incorporate more characteristics. However, they do not identify the characteristics that affect performance in a critical way, and do not incorporate them explicitly in their performance model. In contrast, we propose a systematic procedure to create models that incorporate critical characteristics, aiming at the selection of the best algorithm for solving a given instance. To validate our approach we carried out experiments using an extensive test set. In particular, for the classical bin packing problem, we developed models that incorporate the interrelation among five critical characteristics and the performance of seven heuristic algorithms. As a result of applying our procedure, we obtained a 76% accuracy in the selection of the best algorithm.

Self-Tuning Mechanism for Genetic Algorithms Parameters, an Application to Data-Object Allocation in the Web

In this paper, a new mechanism for automatically obtaining some control parameter values for Genetic Algorithms is presented, which is independent of problem domain and size. This approach differs from the traditional methods which require knowing first the problem domain, and then knowing how to select the parameter values for solving specific problem instances. The proposed method is based on a sample of problem instances, whose solution permits to characterize the problem and to obtain the parameter values.To test the method, a combinatorial optimization model for data-objects allocation in the Web (known as DFAR) was solved using Genetic Algorithms. We show how the proposed mechanism permits to develop a set of mathematical expressions that relates the problem instance size to the control parameters of the algorithm. The experimental results show that the self-tuning of control parameter values of the Genetic Algorithm for a given instance is possible, and that this mechanism yields satisfactory results in quality and execution time. We consider that the proposed method principles can be extended for the self-tuning of control parameters for other heuristic algorithms.

Design of a shared ontology used for translating negotiation primitives

In this paper we present the design of a shared ontology, with the objective to translate a variety of negotiation primitives. Our approach focuses on facilitating communication among agents during negotiation process execution. Traditional negotiation systems impose several restrictions on the type and format of negotiation primitives that can be exchanged among agents. In contrast, we propose the incorporation of an ontology-based solution to overcome heterogeneity and provide communication facilities for participation in negotiations based in open environments such as Internet. To evaluate our ontology we implemented a Web service-oriented negotiation system, and incorporated a translation module that uses the ontology as a vocabulary of negotiation primitives. The experimental results show that the incorporation of the ontology improves the continuity of the execution of negotiation processes, resulting in more agreements.

An ontological approach for translating messages in e-negotiation systems

Traditional negotiation systems have been implemented using agent architectures, where agents communicate through the exchange of messages, based on particular language definitions implicitly encoded, using different implementations and meaning in their messages. Our approach focuses on solving the language heterogeneity problem between agents during a negotiation process, by incorporating an ontology-based translator solution, which is executed only when a misunderstanding occurs. We designed the translator architecture considering that agents involved in a negotiation process may be using similar languages, and not all exchanged messages will cause failures due to misunderstandings. We executed experiments in a Web-based electronic negotiation system, incorporating multiple agents with different language syntax and meaning. The experimental tests show that the proposed solution improves the continuity of the execution of negotiation processes, resulting in more agreements.

Ontology Support for Translating Negotiation Primitives

In this paper we present an ontology solution to solve the problem of language heterogeneity among negotiating agents during the exchange of messages over Internet. Traditional negotiation systems have been implemented using different syntax and semantics. Our proposal offers a novel solution incorporating an ontology, which serves as a shared vocabulary of negotiation messages; and a translation module that is executed on the occurrence of a misunderstanding. We implemented a service oriented architecture for executing negotiations and conducted experiments incorporating different negotiation messages. The results of the tests show that the proposed solution improves the interoperability between heterogeneous negotiation agents.

An Application of Causality for Representing and Providing Formal Explanations about the Behavior of the Threshold Accepting Algorithm

The problem of algorithm selection for solving NP problems arises with the appearance of a variety of heuristic algorithms. The first works claimed the supremacy of some algorithm for a given problem. Subsequent works revealed the supremacy of algorithms only applied to a subset of instances. However, it was not explained why an algorithm solved better a subset of instances. In this respect, this work approaches the problem of explaining through causal model the interrelations between instances characteristics and the inner workings of algorithms. For validating the results of the proposed approach, a set of experiments was carried out in a study case of the Threshold Accepting algorithm to solve the Bin Packing problem. Finally, the proposed approach can be useful for redesigning the logic of heuristic algorithms and for justifying the use of an algorithm to solve an instance subset. This information could contribute to algorithm selection for NP problems.