Catálogo de publicaciones - libros

An improved genetic algorithm is presented to select optimal web services composite plans from a lot of composite plans on the basis of global Quality-of-Service (QoS) constraints. The relation matrix coding scheme of genome is its basis. In this genetic algorithm, an especial fitness function and a mutation policy are proposed on the basis of the relation matrix coding scheme of genome. They enhance convergence of genetic algorithm and can get more excellent composite service plan because they accord with web services selection very well. The simulation results on QoS-aware web services selection have shown that the improved genetic algorithm can gain effectively the composite service plan that satisfies the global QoS requirements, and that the convergence of genetic algorithm was improved very well.

In this paper we propose a framework for modeling mobile information systems. Mobility introduces several challenges and issues that impact the development of mobile systems. As a result, we want applications running on mobile devices to exhibit certain traits; they should be aware of the mobility and be adaptive to the changes that occur due to it. Literature has identified several types of mobility – among them, physical and logical mobility. The former pertains to tangible mobile entities like cars, devices and people, while the latter encompasses mobile software entities. In addition to these, this paper includes the concept of – the movement of a network connection between overlapping networks – in a UML profile for modeling mobile information systems. We discuss our experiences from a case study described in [1] , where we modeled a simple mobile information system and transformed parts of the model into code.

Smart item environments consist of networked nodes with heterogeneous hardware equipment and intermittent network connections. Using a common component technology allows for flexible distribution of components for processing of smart item data. Finding a good deployment plan for a new set of components in an infrastructure is called Component Placement Problem. We propose an approach for finding suitable deployment plans for components with special regard to the characteristics of smart item environments. Our method evaluates deployment plans in terms of both resource consumption and availability. From the analysis of the solution space we conclude that the number of network link uses is an important criterion for the quality of a deployment plan regarding both cost and availability. Based on this finding, we have derived a heuristic that creates deployment plans, which have a low number of link uses and are thus more likely of high quality.

Dynamic loading of code is a crucial and often neglected part of today’s distributed systems that face increasing dynamics, complexity and heterogeneity. Ubiquitous computing and mobile computing even strengthen this trend. As the local availability of suitable code cannot be assumed in such environments, we propose a generic, decentralised code loading infrastructure. The whole process of publication, look-up, implementation selection and the final loading of platform-specific code is decentralised and requires only basic peer-to-peer functionality. In contrast to previous work, our infrastructure allows any peer participating in the network to offer and to obtain platform-specific code in a dynamic and heterogeneous environment. By building on our generic concept, we present a JXTA-based service for dynamic code loading, which is realised by extending and improving JXTA-built-in mechanisms for dynamic service integration. Subsequently, we show the practical application of our infrastructure by an integration into our CORBA middleware and an implementation of mobile objects and mobile web services.