Francisco Gortázar

Hybrid scatter tabu search for unconstrained global optimization

The problem of finding a global optimum of an unconstrained multimodal function has been the subject of intensive study in recent years, giving rise to valuable advances in solution methods. We examine this problem within the framework of adaptive memory programming (AMP), focusing particularly on AMP strategies that derive from an integration of Scatter Search and Tabu Search. Computational comparisons involving 16 leading methods for multimodal function optimization, performed on a testbed of 64 problems widely used to calibrate the performance of such methods, disclose that our new Scatter Tabu Search (STS) procedure is competitive with the state-of-the-art methods in terms of the average optimality gap achieved.

Path relinking for large-scale global optimization

In this paper we consider the problem of finding a global optimum of a multimodal function applying path relinking. In particular, we target unconstrained large-scale problems and compare two variants of this methodology: the static and the evolutionary path relinking (EvoPR). Both are based on the strategy of creating trajectories of moves passing through high-quality solutions in order to incorporate their attributes to the explored solutions. Computational comparisons are performed on a test-bed of 19 global optimization functions previously reported with dimensions ranging from 50 to 1,000, totalizing 95 instances. Our results show that the EvoPR procedure is competitive with the state-of-the-art methods in terms of the average optimality gap achieved. Statistical analysis is applied to draw significant conclusions.

Black box scatter search for general classes of binary optimization problems

The purpose of this paper is to apply the scatter search methodology to general classes of binary problems. We focus on optimization problems for which the solutions are represented as binary vectors and that may or may not include constraints. Binary problems arise in a variety of settings, including engineering design and statistical mechanics (e.g., the spin glass problem). A distinction is made between two sets of general constraint types that are handled directly by the solver and other constraints that are addressed via penalty functions. In both cases, however, the heuristic treats the objective function evaluation as a black box. We perform computational experiments with four well-known binary optimization problems to study the efficiency (speed) and effectiveness (solution quality) of the proposed method. Comparisons are made against both commercial software and specialized procedures on a set of 376 instances. We chose commercial software that is similar in nature to the proposed procedure, namely, it treats the objective function as a black box and the search is based on evolutionary optimization techniques.

A black-box scatter search for optimization problems with integer variables

The goal of this work is the development of a black-box solver based on the scatter search methodology. In particular, we seek a solver capable of obtaining high quality outcomes to optimization problems for which solutions are represented as a vector of integer values. We refer to these problems as integer optimization problems. We assume that the decision variables are bounded and that there may be constraints that require that the black-box evaluator is called in order to know whether they are satisfied. Problems of this type are common in operational research areas of applications such as telecommunications, project management, engineering design and the like.Our experimental testing includes 171 instances within four classes of problems taken from the literature. The experiments compare the performance of the proposed method with both the best context-specific procedures designed for each class of problem as well as context-independent commercial software. The experiments show that the proposed solution method competes well against commercial software and that can be competitive with specialized procedures in some problem classes.

A hybrid metaheuristic for the cyclic antibandwidth problem

We propose a hybrid artificial bee colony algorithm for the cyclic antibandwidth problem.We present a computational comparison of different parameter settings.We derive a fine-tuning hybrid artificial bee colony algorithm.The proposal is very competitive with the state-of-the-art algorithm for the cyclic antibandwidth problem. In this paper, we propose a hybrid metaheuristic algorithm to solve the cyclic antibandwidth problem. This hard optimization problem consists of embedding an n-vertex graph into the cycle Cn, such that the minimum distance (measured in the cycle) of adjacent vertices is maximized. It constitutes a natural extension of the well-known antibandwidth problem, and can be viewed as the dual problem of the cyclic bandwidth problem.Our method hybridizes the artificial bee colony methodology with tabu search to obtain high-quality solutions in short computational times. Artificial bee colony is a recent swarm intelligence technique based on the intelligent foraging behavior of honeybees. The performance of this algorithm is basically determined by two search strategies, an initialization scheme that is employed to construct initial solutions and a method for generating neighboring solutions. On the other hand, tabu search is an adaptive memory programming methodology introduced in the eighties to solve hard combinatorial optimization problems. Our hybrid approach adapts some elements of both methodologies, artificial bee colony and tabu search, to the cyclic antibandwidth problem. In addition, it incorporates a fast local search procedure to enhance the local intensification capability. Through the analysis of experimental results, the highly effective performance of the proposed algorithm is shown with respect to the current state-of-the-art algorithm for this problem.

Variable neighborhood search with ejection chains for the antibandwidth problem

In this paper, we address the optimization problem arising in some practical applications in which we want to maximize the minimum difference between the labels of adjacent elements. For example, in the context of location models, the elements can represent sensitive facilities or chemicals and their labels locations, and the objective is to locate (label) them in a way that avoids placing some of them too close together (since it can be risky). This optimization problem is referred to as the antibandwidth maximization problem (AMP) and, modeled in terms of graphs, consists of labeling the vertices with different integers or labels such that the minimum difference between the labels of adjacent vertices is maximized. This optimization problem is the dual of the well-known bandwidth problem and it is also known as the separation problem or directly as the dual bandwidth problem. In this paper, we first review the previous methods for the AMP and then propose a heuristic algorithm based on the variable neighborhood search methodology to obtain high quality solutions. One of our neighborhoods implements ejection chains which have been successfully applied in the context of tabu search. Our extensive experimentation with 236 previously reported instances shows that the proposed procedure outperforms existing methods in terms of solution quality.

Tabu search for the Max-Mean Dispersion Problem

In this paper, we address a variant of a classical optimization model in the context of maximizing the diversity of a set of elements. In particular, we propose heuristics to maximize the mean dispersion of the selected elements in a given set. This NP-hard problem was recently introduced as the maximum mean dispersion problem (MaxMeanDP), and it models several real problems, from pollution control to ranking of web pages. In this paper, we first review the previous methods for the MaxMeanDP, and then explore different tabu search approaches, and their influence on the quality of the solutions obtained. As a result, we propose a dynamic tabu search algorithm, based on three different neighborhoods. Experiments on previously reported instances show that the proposed procedure outperforms existing methods in terms of solution quality. It must be noted that our findings on the use of different memory structures invite to further consideration of the interplay between short and long term memory to enhance simple forms of tabu search.

Heuristics for the bandwidth colouring problem

The bandwidth colouring problem consists of assigning a colour to each vertex of a graph, so that the absolute value of the difference between the colours of adjacent vertices is at least the value of the weight of the associated edge. This problem generalises the classical vertex colouring problem and different heuristics have recently been proposed to obtain high quality solutions. In this paper we describe both memory-based and memory-less methods to solve the bandwidth colouring problem. In particular we propose new constructive and improvement methods based on tabu search and GRASP. Comparison of our results with previously reported instances and existing heuristics indicate that the methods we propose are competitive and require short computational times. Our findings also disclose that memory appears to play a more important role during improvement phases of search than during constructive phases.

GRASP with path relinking for the single row facility layout problem

The single row facility layout problem (SRFLP) is an NP -hard problem that consists of finding an optimal arrangement of a set of rectangular facilities (with equal height and different lengths), placing them next to each other along a line. The SRFLP has practical applications in contexts such as arranging rooms along corridors, setting books on shelves, allocating information on magnetic disks, storing items in warehouses, or designing layouts for machines in manufacturing systems. This paper combines the greedy randomized adaptive search procedure (GRASP) methodology, and path relinking (PR) in order to efficiently search for high-quality solutions for the SRFLP. In particular, we introduce: (i) several construction procedures, (ii) a new fast local search strategy, and (iii) an approach related to the Ulam distance in order to construct short path relinking trajectories. We also present a new set of large challenging instances, since previous sets do not allow to determine significant differences among advanced metaheuristics. Experiments show that our procedure outperforms state-of-the-art methods in all of the scenarios we considered. Firstly, the GRASP with PR finds the best known solutions for previous instances used in the literature, but employing considerably less computing time than its competitors. Secondly, our method outperforms the current state-of-the-art methods in 38 out of 40 new instances when running for the same amount of computing time. Finally, nonparametric tests for detecting differences between algorithms report p-values below 10 - 11 , which supports the superiority of our approach.

WebRTC Testing: Challenges and Practical Solutions

WebRTC comprises a set of novel technologies and standards that provide Real-Time Communication on Web browsers. WebRTC makes simple the embedding of voice and video communications in all types of applications. However, releasing those applications to production is still very challenging due to the complexity of their testing. Validating a WebRTC service requires assessing many functional (e.g. signaling logic, media connectivity, etc.) and non-functional (e.g. quality of experience, interoperability, scalability, etc.) properties on large, complex, distributed and heterogeneous systems that spawn across client devices, networks and cloud infrastructures. In this article, we present a novel methodology and an associated tool for doing it at scale and in an automated way. Our strategy is based on a blackbox end-to-end approach through which we use an automated containerized cloud environment for instrumenting Web browser clients, which benchmark the SUT (system under test), and fake clients, that load it. Through these benchmarks, we obtain, in a reliable and statistically significant way, both network-dependent QoS (Quality of Service) metrics and media-dependent QoE (Quality of Experience) indicators. These are fed, at a second stage, to a number of testing assertions that validate the appropriateness of the functional and non-functional properties of the SUT under controlled and configurable load and fail conditions. To finish, we illustrate our experiences using such tool and methodology in the context of the Kurento open source software project and conclude that they are suitable for validating large and complex WebRTC systems at scale.