Jose A. Mocholi

Strategies for accelerating ant colony optimization algorithms on graphical processing units

Ant colony optimization (ACO) is being used to solve many combinatorial problems. However, existing implementations fail to solve large instances of problems effectively. In this paper we propose two ACO implementations that use graphical processing units to support the needed computation. We also provide experimental results by solving several instances of the well-known orienteering problem to show their features, emphasizing the good properties that make these implementations extremely competitive versus parallel approaches.

A Metamodeling Approach for Requirements Specification

There are many Requirements Engineering approaches and techniques that help to specify, analyze and validate requirements. However, they are neither widely accepted nor widely used by the industrial software community. One of the main problems faced when applying them is to what extent they can be easily adapted to the specific needs of the project. This has often led to unsatisfactory requirements management in industrial software development. In this work, a proposal for requirements modeling is presented that allows the integration of the expressiveness of some of the more relevant Requirements Engineering techniques by taking advantage of metamodeling. This proposal focuses on scalability with respect to expressiveness and adaptability to the application domain to establish some basic guidelines and extension mechanisms that lend coherence and semantic precision. A case study and 4 tool support are presented to describe the application of this proposal in a real-life project.

AGORAS: Exploring Creative Learning on Tangible User Interfaces

Departing from creative learning foundations, this paper discusses on the suitability of interactive tables as a grounding technology to support creative learning for several reasons: support for social learning, because the subjects share a physical space as in traditional non-digital technologies, communication during the creative, experimental and reflexive process is direct and not computer-mediated, and subjects can carry out the task in parallel on the same surface. Considering reflection, discussion and creation processes in a loop, an experiment with teenagers has been conducted comparing a digital-based against a pure tangible tabletop in a task of creating entities consisting of blocks and joint elements. This preliminary study, designed to obtain initial insights about whether the grounding technology may become a promising tool to support creative learning, explores some aspects such as productivity, complexity of designs and concurrent co-manipulation. The results showed that subjects were more productive in terms of the number of solutions obtained using the non computer-mediated approach. However using the digital tabletop approach subjects design, on average, more complex or elaborate solutions in terms of the number of involved bodies and joints. Finally, an important finding was that teams established more frequently concurrent cooperation schemes in the digital tabletop condition by sharing more effectively the creation space.

An emotionally biased ant colony algorithm for pathfinding in games

Pathfinding is one of the tasks, apart from graphics rendering, requiring most CPU resources. Although there are many approaches to effectively solve pathfinding problems, they are becoming less suitable as more and more games have larger game worlds that dynamically change during the game play. These new games have more visually realistic graphics that increase the game characters realism but all these efforts may be useless if game characters perform dumb movements or follow inappropriate paths such as repeatedly walking close to an enemy or a predator while moving from one location to another. To tackle this problem we present in this paper an ant colony algorithm for path finding that takes into account the emotions of the game characters and we show how our approach is used in an augmented-reality educational game. The proposed algorithm is implemented on a GPU processor to demonstrate its scalability with large problem sizes when compared to its corresponding CPU version.

Ant colony optimisation for resource searching in dynamic peer-to-peer grids

The applicability of peer-to-peer p2p in the domain of grid computing has been an important subject over the past years. Nevertheless, the sole merger between p2p and the concept of grid is not sufficient to guarantee non-trivial efficiency. Some claim that ant colony optimisation ACO algorithms might provide a definite answer to this question. However, the use of ACO in grid networks causes several problems. The first and foremost stems out of the fact that ACO algorithms usually perform well under the conditions of static networks, solving predetermined problems in a known and bound space. The question that remains to be answered is whether the evolutive component of these algorithms is able to cope with changing conditions; and by those we mean changes both in the positive sense, such as the appearance of new resources, but also in the negative sense, such as the disappearance or failure of fragments of the network. In this paper we study these considerations in depth, bearing in mind the specificity of the peer-to-peer nature.

A multicriteria ant colony algorithm for generating music playlists

Highlights? We address the problem of music playlist generation based on the user- personalized specification of context information. ? We designed an ACO algorithm that is able to tackle both the optimization problem and the personalization requirement. ? Experiments show the algorithm is able to create good playlists while observing the personalization requirements. ? The algorithm can be applied to other domains by using the appropriate ontology and defining evaluation functions. In this paper we address the problem of music playlist generation based on the user-personalized specification of context information. We propose a generic semantic multicriteria ant colony algorithm capable of dealing with domain-specific problems by the use of ontologies. It also employs any associated metadata defined in the search space to feed its solution-building process and considers any restrictions the user may have specified. An example is given of the use of the algorithm for the problem of automatic generation of music playlists, some experimental results are presented and the behavior of the approach is explained in different situations.

A meta-model for dataflow-based rules in smart environments: Evaluating user comprehension and performance

A considerable part of the behavior in smart environments relies on event-driven and rule specification. Rules are the mechanism most often used to enable user customization of the environment. However, the expressiveness of the rules available to users in editing and other tools is usually either limited or the available rule editing interfaces are not designed for end-users with low skills in programming. This means we have to look for interaction techniques and new ways to define user customization rules. This paper describes a generic and flexible meta-model to support expressive rules enhanced with data flow expressions that will graphically support the definition of rules without writing code. An empirical study was conducted on the ease of understanding of the visual data flow expressions, which are the key elements in our rule proposal. The visual dataflow language was compared to its corresponding textual version in terms of comprehension and ease of learning by teenagers in exercises involving calculations, modifications, writing and detecting equivalences in expressions in both languages. Although the subjects had some previous experience in editing mathematical expressions on spreadsheets, the study found their performance with visual dataflows to be significantly better in calculation and modification exercises. This makes our dataflow approach a promising mechanism for expressing user-customized reactive behavior in Ambient Intelligence (AmI) environments. The performance of the rule matching processor was validated by means of two stress tests to ensure that the meta-model approach adopted would be able to scale up with the number of types and instances in the space.

Learning semantically-annotated routes for context-aware recommendations on map navigation systems

Modern technology has brought many changes to our everyday lives. Our need to be in constant touch with others has been met with the cellphone, which has become our companion and the convergence point of many technological advances. The combination of capabilities such as browsing the Internet and GPS reception has multiplied the services and applications based on the current location of the user. However, providing the user with these services has certain drawbacks. Although map navigation systems are the most meaningful way of displaying this information, the user still has to manually set up the filter in order to obtain a non-bloated visualization of the map and the available services. To tackle this problem, we present here a semantic multicriteria ant colony algorithm capable of learning the users routes, including associated context information, and then predicting the most likely route a user is following, given his current location and context data. This knowledge could then be used as the basis for offering services related to his current (or most likely future) context data close to the path he is following. Our experimental results show that our algorithm is capable of obtaining consistent solutions sets even when multiple objective ontological terms are included in the process.

On the performance of ACO-based methods in p2p resource discovery

Over the recent years peer-to-peer (p2p) systems have become increasingly popular. As of today most of the internet IP traffic is already transmitted in this format and still it is said to double in volume till 2014. Most p2p systems, however, are not pure serverless solutions, nor is the searching in those networks highly efficient, usually achieved by simple flooding. In order to confront with the growing traffic we must consider more elaborate search mechanisms and far less centralized environments. An effective proposal to this problem is to solve it in the domain of ant colony optimization metaheuristics. In this paper we present an overview of ACO algorithms that offer the best potential in this field, under the strict requirements and limitations of a pure p2p network. We design several experiments to serve as an evaluation platform for the mentioned algorithms to conclude the features of a high quality approach. Finally, we consider two hybrid extensions to the classical algorithms, in order to examine their contribution to the overall quality robustness.

TangiWheel: A Widget for Manipulating Collections on Tabletop Displays Supporting Hybrid Input Modality

In this paper we present TangiWheel, a collection manipulation widget for tabletop displays. Our implementation is flexible, allowing either multi-touch or interaction, or even a hybrid scheme to better suit user choice and convenience. Different TangiWheel aspects and features are compared with other existing widgets for collection manipulation. The study reveals that TangiWheel is the first proposal to support a hybrid input modality with large resemblance levels between touch and tangible interaction styles. Several experiments were conducted to evaluate the techniques used in each input scheme for a better understanding of tangible surface interfaces in complex tasks performed by a single user (e.g., involving a typical master-slave exploration pattern). The results show that tangibles perform significantly better than fingers, despite dealing with a greater number of interactions, in situations that require a large number of acquisitions and basic manipulation tasks such as establishing location and orientation. However, when users have to perform multiple exploration and selection operations that do not require previous basic manipulation tasks, for instance when collections are fixed in the interface layout, touch input is significantly better in terms of required time and number of actions. Finally, when a more elastic collection layout or more complex additional insertion or displacement operations are needed, the hybrid and tangible approaches clearly outperform finger-based interactions.