Yves Vanrompay

Genetic algorithm-based optimization of service composition and deployment

Services running on mobile systems must be able to adapt themselves to changing user needs and availability of resources. We propose to use Genetic Algorithms to search for the best service variant in the current context. The chosen service composition is then deployed on a set of available nodes in an optimal way. We illustrate that Genetic Algorithms provide a scalable and self-organizing solution to service composition and deployment. We argue that the approach meets some main requirements demanded by services running on mobile systems. A motivating scenario is presented in which a distributed server allows users to share content and run applications in mobile ad-hoc networks.

Middleware for the Internet of Things, design goals and challenges

As the number of wireless devices increases and their size becomes smaller, there can be more interaction between everyday objects of our life. With advances in RFID chips and the introduction of new generations of these devices that are smaller and cheaper, it is possible to put a wireless interface on almost all everyday objects: vehicles, clothes, foodstuffs, etc. This concept is called the \textit{Internet of Things}. Interaction with thousands of wireless devices leads to a continuous and massive flow of events which are generated spontaneously. The question of how to deal with this enormous number of events is challenging and introduces new design goals for a communication mechanism. In this paper we argue that a middleware together with suitable linguistic abstractions is a proper solution. We also point out the challenges in developing this middleware. Moreover, we give an overview of recent related work and describe why they fail to address these challenges.

Context inference of users' social relationships and distributed policy management

Inference of high-level context is becoming crucial in development of context-aware applications. An example is social context inference - i.e., deriving social relations based upon the users daily communication with other people. The efficiency of this mechanism mainly depends on the method(s) used to draw inferences based on existing evidence and sample information, such as a training data. Our approach uses rule-based data mining, Bayesian network inference, and user feedback to compute the probabilities of another user being in the specific social relationship with a user whose daily communication is logged by a mobile phone. In addition, a privacy mechanism is required to ensure the users personal integrity and privacy when sharing this users sensitive context data. Therefore, the derived social relations are used to define a users policies for context access control, which grant the restricted context information scope depending on the users current context. Finally, we propose a distributed architecture capable of managing this context information based upon these context access policies.

Analyzing the efficiency of context-based grouping on collaboration in VANETs with large-scale simulation

Vehicle-to-vehicle and vehicle-to-infrastructure communication systems enable vehicles to share information captured by their local sensors with other interested vehicles. To ensure that this information is delivered at the right time and location, context-aware routing is vital for intelligent inter-vehicular communication. Traditional network addressing and routing schemes do not scale well for large vehicular networks. The conventional network multicasting and broadcasting cause significant overhead due to a large amount of irrelevant and redundant transmissions. To address these challenges, we first take into account contextual properties such as location, direction, and information interest to reduce the network traffic overhead. Second, to improve the relevancy of the received information we leverage the mobility patterns of vehicles and the road layouts to further optimize the peer-to-peer routing of the information. Third, to ensure our approach is scalable, we propose a context-based grouping mechanism in which relevant information is shared in an intelligent way within and between the groups. We evaluate our approach based on groups with common spatio-temporal characteristics. Our simulation experiments show that our context-based routing scheme and grouping mechanism significantly reduces the propagation of irrelevant and redundant information.

Optimizing information dissemination in large scale mobile peer-to-peer networks using context-based grouping

Nowadays novel embedded computing devices enable vehicles to form large scale mobile peer-to-peer networks in which they can assist each other to improve their driving experience. Therefore context-aware communication is considered to be vital for inducing inter-vehicular intelligence between groups of vehicles with similar interests. However, traditional network addressing schemes are not well suited for group-based communication in large scale vehicular networks. The classical network paradigms of multicasting and broadcasting to define groups are too limited. First, there is no way to optimize network traffic based on the contextual characteristics of the nodes. Second, the groups of nodes are highly dynamic with vehicles randomly joining and leaving multiple groups. We propose an information dissemination approach based on context grouping in which only relevant information is shared among nodes. We evaluate our approach in a large scale vehicular network where groups are formed based on the location and shared interests of the nodes. The experiments show that by inducing our context-based grouping mechanism we can significantly eliminate irrelevant information and reduce overall network traffic in a scalable way.

Context Grouping Mechanism for Context Distribution in Ubiquitous Environments

Context distribution is a key aspect for successful applications within mobile and ubiquitous computing environments. In such environments, context information is acquired by several and multiple context sensors distributed over the environment. Applications collect and react to these data, according to predefined adaptation mechanisms. The success of these mechanisms depends on the availability of context information, which is disseminated over the network. However, in practice, only a fraction of the observable context information is required by the adaptation mechanisms. Moreover, for privacy reasons, it is important to delimitate a scope for context dissemination. In this work we address these issues by proposing a context grouping mechanism which allows the definition of groups based on the context characteristics. Each group is defined by these characteristics and delimitate a given context information set that can be distributed among group members. This approach of context grouping acts as a two-fold mechanism. On the one hand, it controls and organizes context distribution over a peer-to-peer network. On the other hand, it proposes a primary and low-level privacy mechanism for context distribution, which is an important aspect influencing context distribution.

A peer-to-peer based infrastructure for context distribution in mobile and ubiquitous environments

The increasing proliferation of mobile devices has raised the expectations for user-customized and environment-aware services. However, mobile context-aware systems inherently feature characteristics of distribution and heterogeneity which pose great challenges to their developers. This paper focuses on context distribution in mobile and ubiquitous computing environments. We propose a peer-to-peer based context distribution approach and evaluate it against the derived requirements.

Fadip: lightweight publish/subscribe for mobile ad hoc networks

Publish/Subscribe is an interesting communication paradigm because it fosters a high degree of decoupling between the communicating parties and provides the ability to communicate in an asynchronous way. Publish/Subscribe systems have been extensively studied for wired networks but designing a Publish/Subscribe system for mobile ad hoc networks is still a challenge. In this paper we propose a lightweight Publish/ Subscribe system for mobile ad hoc networks which uses a limited gossip mechanism to match the published messages with the subscriptions. The goal of this work is to reduce the number of exchanged messages used for communication and maintenance while keeping an acceptable delivery ratio. Experimental results show that Fadip achieves an acceptable delivery ratio even in high mobility rates.

Context-Aware Service Selection with Uncertain Context Information

The current evolution of Service-Oriented Computing in ubiquitous systems is leading to the development of context-aware services. These are services whose description is enriched with context information related to the service execution environment and adaptation capabilities. This information is often used for discovery and adaptation purposes. However, in real-life systems context information is naturally dynamic, uncertain and incomplete, which represents an important issue when comparing service description and user requirements. Uncertainty of context information may lead to an inexact match between provided and required service capabilities, and consequently to the non-selection of services. In order to handle uncertain and incomplete context information, we propose a mechanism inspired by graph-comparison for matching contextual service descriptions using similarity measures that allow inexact matching. Service description and requirements are compared using two kinds of similarity measures: local measures, which compare individually required and provided properties, and global measures, which take into account the context description as a whole. We show how the proposed mechanism is integrated in MUSIC, an existing adaptation middleware, and how it enables more optimal adaptation decision making.

Evolution of context-aware user profiles

Context-awareness and adaptation are key issues in mobile and ubiquitous computing. Applications on mobile devices use context information to adapt themselves to changing environments. User profiles play an important role in these systems as they serve as an individualization filter in a wide range of possible context adaptation parameters. In this paper we propose a modeling approach for the evolution of context-aware user profiles. A motivating scenario, the intelligent selection of a suitable medical expert in an emergency situation, shows the need for context-aware matching of user profiles. This is achieved by a similarity matching algorithm and reinforcement learning.