Jos Nijhuis

COVAMOF: A framework for modeling variability in software product families

A key aspect of variability management in software product families is the explicit representation of the variability. Experiences at several industrial software development companies have shown that a software variability model should do four things: (1) uniformly represent variation points as first-class entities in all abstraction layers (ranging from features to code), (2) allow for the hierarchical organization of the variability, (3) allow for the first-class representation of simple (i.e., one-to-one) and complex (i.e., n-to-m) dependencies, and (4) allow for modeling the relations between dependencies. Existing variability modeling approaches support the first two requirements, but lack support for the latter two. The contribution of this paper is a framework for variability modeling—COVAMOF—that provides support for all four requirements.

Design Decisions: The Bridge between Rationale and Architecture

Software architecture can be seen as a decision making process; it involves making the right decisions at the right time. Typically, these design decisions are not explicitly represented in the artifacts describing the design. They reside in the minds of the designers and are therefore easily lost. Rationale management is often proposed as a solution, but lacks a close relationship with software architecture artifacts. Explicit modeling of design decisions in the software architecture bridges this gap, as it allows for a close integration of rationale management with software architecture. This improves the understandability of the software architecture. Consequently, the software architecture becomes easier to communicate, maintain and evolve. Furthermore, it allows for analysis, improvement, and reuse of design decisions in the design process.

Modeling dependencies in product families with COVAMOF

Many variability modeling approaches consider only formalized dependencies, i.e. in- or exclude relations between variants. However, in real industrial product families, dependencies are often much more complicated. In this paper, we discuss the product derivation problems associated with dependencies, and show how our variability modeling framework COVAMOF addresses these issues. Throughout the paper, we use examples of Intrada, an intelligent traffic systems family of Dacolian B.V

Neurocontrol for lateral vehicle guidance

A solution to autonomous lateral vehicle guidance using a neurocontroller that can learn from measured human-driving data without knowledge of the physical car parameters is discussed. Simulations and practical tests confirm that a small-size feedforward autonomous neural network (21 neurons) can learn to steer a vehicle at high speeds only from looking at human-driving examples. In this way, the network learns the total closed-loop behavior, including the nonlinear dynamics of the vehicle and the drivers individual driving style. The main result of practical investigations is that the neutral controller trained on human-driving examples exhibits an aperiodic behavior that does not vanish at higher speeds (tests performed up to 130 km/h) and produces fewer lateral deviations than the linear state controller.<<ETX>>

DySOA: making service systems self-adaptive

Service-centric systems exist in a very dynamic environment. This requires these systems to adapt at runtime in order to keep fulfilling their QoS. In order to create self-adaptive service systems, developers should not only design the service architecture, but also need to design the self-adaptability aspects in a structured way. A key aspect in creating these self-adaptive service systems is modeling runtime variability properties. In this paper, we propose DySOA (Dynamic Service-Oriented Architecture), an architecture that extends service-centric applications to make them self-adaptive. DySOA allows developers to explicitly model elements that deal with QoS evaluation and variable composition configurations. Having the DySOA elements explicit enables separation of concerns, making them adaptable at runtime and reusable in next versions. We demonstrate the use of DySOA with an example.

COSVAM: a technique for assessing software variability in software product families

Evolution of variability is a key factor in the successful exploitation of commonalities in software product families. Assessment of variability can be used to determine how the variability provided by a product family should evolve. We present COSVAM (COVAMOF software variability assessment method), a variability assessment technique that specifically addresses evolution of variability. We exemplify our approach with the Dacolian case study.

Structure and application of NNSIM: a general-purpose neural network SIMulator

Abstract The paper describes the neural network simulator NNSIM. The procedural interface provides a very flexible way to simulate a neural network application in a non-neural environment. To comfort the evaluttion of neural network properties, an interruptive mode is supported. This interruptive change of network parameters without compiling or loading the network anew allows for the automated optimization of the network topology and its parameters. The functionality of NNSIM is illustrated in the design of a mixed neural-digital image processing system.

Transformational DT-CNN design from morphological specifications

Morphology provides the algebraic means to specify operations on images. Discrete-time cellular neural networks (DT-CNNs) mechanize the execution of operations on images. The paper first shows the equivalence between morphological functions and DT-CNNs. Then, the argument is extended to the synthesis of optimal DT-CNN structures from complex morphological expressions. It is shown that morphological specifications may be freely derived, to be subsequently transformed and adopted to the needs of a specific target terminology. This process of technology mapping can be automated along the well-trodden path in CAD for microelectronics.

A context architecture for service-centric systems

Service-centric systems are highly dynamic and often complex systems, with different services running on a distributed network. For the design of context-aware service-centric systems, paradigms have to be developed that deal with the distributed and dynamic nature of these systems, and with the unreliability and unavailability problems of providing information on their context. This paper presents a context architecture for the development of context-aware service-centric systems that provides the context information and deals with these challenges.

Neural network approaches to capture temporal information

The automated design and construction of neural networks receives growing attention of the neural networks community. Both the growing availability of computing power and development of mathematical and probabilistic theory have had severe impact on the design and modelling approaches of neural networks. This impact is most apparent in the use of neural networks to time series prediction. In this paper, we give our views on past, contemporary and future design and modelling approaches to neural forecasting.