Antonio Gonzalez-Pardo

A new CSP graph-based representation for Ant Colony Optimization

Constraint Satisfaction Problems (CSP) have been widely studied in several research areas like Artificial Intelligence or Operational Research due their complexity and industrial interest. From previous research areas, heuristic (informed) search methods have been particularly active looking for feasible approaches. One of the critical problems to work with CSP is related to the exponential growth of computational resources needed to solve even the simplest problems. This paper presents a new efficient CSP graph-based representation to solve CSP by using Ant Colony Optimization (ACO) algorithms. This paper presents also a new heuristic (called Oblivion Rate), that have been designed to improve the current state-of-the-art in the application of ACO algorithms on these domains. The presented graph construction provides a strong reduction in both, the number of connections and the number of nodes needed to model the CSP. Also, the new heuristic is used to reduce the number of pheromones in the system (allowing to solve problems with an increasing complexity). This new approach has been tested, as case study, using the classical N-Queens Problem. Experimental results show how the new approach works in both, reducing the complexity of the resulting CSP graph and solving problems with increasing complexity through the utilization of the Oblivion Rate.

A new CSP graph-based representation to resource-constrained project scheduling problem

Resource-Constrained Project Scheduling Problem (RCPSP) is a NP-hard combinatorial problem that consists in scheduling different activities in such a way the resource, precedence, and temporal constraints are satisfied. The main problem when dealing with NP-hard problems is the exponential growth of the computational resources needed to solve the problems. This work is an extension of a previous one, where a new CSP graph-based representation to solve Constraint Satisfaction Problems (CSP) by using Ant Colony Optimization (ACO) were proposed. This paper studies the behaviour of the CSP graph-based representation when it is applied to a real-world complex problem, in this case the RCPSP. The dataset used in this work has been extracted from Project Scheduling Problem Library (PSPLIB). Experimental results show that the proposed approach provides excellent results, closer to the optimum values published in the PSPLIB repository. Also, it has been analysed how the number of jobs and the number of different execution modes affect the performance of the algorithm.

Behaviour-based identification of student communities in virtual worlds

This work has been funded by the Spanish Ministry of Science and Innovation under the project ABANT (TIN2010-19872/TSI).

Using virtual worlds for behaviour clustering-based analysis

Virtual Worlds have become a very popular domain and its high inmersive characteristics can be used to extract information about the avatars behaviour. In this kind of environment it is possible to obtain interesting data about avatars, such as their exact position in the world, what they are looking at (eye-gazing) or what they are talking about. This paper studies how this information, obtained from avatars interactions, can be integrated in order to apply clustering techniques. Monitoring avatars in a virtual world is a useful task that allows the identification of behavioral groups. The meaning of these groups depends on the application domain, for example in educational virtual worlds, they can represent whether students are paying attention to the teachers explanation or not.

Micro and Macro Lemmings Simulations Based on Ants Colonies

Ant Colony Optimization (ACO) has been successfully applied to a wide number of complex and real domains. From classical optimization problems to video games, these kind of swarm-based approaches have been adapted, to be later used, to search for new meta-heuristic based solutions. This paper presents a simple ACO algorithm that uses a specifically designed heuristic, called common-sense, which has been applied in the classical video game Lemmings. In this game a set of lemmings must reach the exit point of each level, using a subset of finite number of skills, taking into account the contextual information given from the level. The paper describes both the graph model and the context-based heuristic, designed to implement our ACO approach. Afterwards, two different kind of simulations have been carried out to analyse the behaviour of the ACO algorithm. On the one hand, a micro simulation, where each ant is used to model a lemming, and a macro simulation where a swarm of lemmings is represented using only one ant. Using both kind of simulations, a complete experimental comparison based on the number and quality of solutions found and the levels solved, is carried out to study the behaviour of the algorithm under different game configurations.

A multi-agent traffic simulation framework for evaluating the impact of traffic lights

This is an electronic version of the paper presented at the 3rd International Conference on Agents and Artificial Intelligence, held in Rome on 2011

A Case Study on Grammatical-Based Representation for Regular Expression Evolution

Regular expressions, or simply regex, have been widely used as a powerful pattern matching and text extractor tool through decades. Although they provide a powerful and flexible notation to define and retrieve patterns from text, the syntax and the grammatical rules of these regex notations are not easy to use, and even to understand. Any regex can be represented as a Deterministic or Non-Deterministic Finite Automata; so it is possible to design a representation to automatically build a regex, and a optimization algorithm able to find the best regex in terms of complexity. This paper introduces both, a graph-based representation for regex, and a particular heuristic-based evolutionary computing algorithm based on grammatical features from this language in a particular data extraction problem.

Optimal Message Interchange in a Self-organizing Multi-agent System

Over the last decade there has been a growing interest on Intelligent Agents and Multi-Agent Systems (MAS) in several fields such as Artificial Intelligence (AI), Software Engineering, Psychology, etc... Different problems can be solved in these fields by creating societies of agents that communicate with each other. Nevertheless, when the number of agents is large and the connectivity is extensive, the system suffers from overhead in the communication among agents due to the large number messages exchanged. This work addresses the search for an optimal communication topology to avoid these situations. This optimal topology is characterized by the use of a redirecting probability in the communication. The redirection of a communication is performed before the execution of the MAS. Once agents start the execution, the topology is fixed and remains unchanged. This characteristic is useful in those systems where a given topology can not be changed as, for example, in wired networks. On the other hand, in the proposed solution agents contain a local message discrimination process as a function of the sender of the message. Experiments show an important improvement in terms of a reduction in the number of iterations needed to solve the problem and also in the number of messages exchanged.

A method for building predictive HSMMs in interactive environments

The study of user behavior based on his/her interactions with a system is widely extended over several fields of research. Often, it is useful to have an underlying model to generate behavioral predictions, allowing the system to automatically adapt to the user and to detect deviations from an expected behavior. In this work, we develop a general method to create, select and validate a Hidden Semi-Markov Model (HSMM) to predict behavior in interactive environments, based on previously seen interactions. The method is completely data-driven, unrestricted by any prior knowledge of the model structure, and easy to automate once some parameters has been adjusted. To test the proposed method, a multi-UAV mission simulator has been used, obtaining a model able to perform adequate predictions in terms of quality and time.

Modeling the Behavior of Unskilled Users in a Multi-UAV Simulation Environment

The use of Unmanned Aerial Vehicles (UAVs) has been growing over the last few years. The accelerated evolution of these systems is generating a high demand of qualified operators, which requires to redesign the training process and focus on a wider range of candidates, including inexperienced users in the field, in order to detect skilled-potential operators. This paper uses data from the interactions of multiple unskilled users in a simple multi-UAV simulator to create a behavioral model through the use of Hidden Markov Models (HMMs). An optimal HMM is validated and analyzed to extract common behavioral patterns among these users, so that it is proven that the model represents correctly the novelty of the users and may be used to detect and predict behaviors in multi-UAV systems.