Damir Bilic

Simulations have long been part of hardware-centric system domains. Similarly, architecture design is a common practice for complex industrial systems, which comprise many components that can be arranged in different layouts according to given requirements. Configuring simulation models and choosing the architecture design can be time-consuming activities. This paper presents a model-driven approach to automate the simulation configuration and architecture layouting engineering activities by leveraging model-driven optimisation techniques. The approach leverages a research solution, MOMoT (Marrying Optimisation and Model Transformations), an academic tool that combines search-based algorithms and model transformations. MOMoT is extended with two software modules, leveraging the Functional Mock-up Interface standard for simulation configuration and an architectural description language to design architecture layouts. Our solution is presented in the context of Volvo Construction Equipment's industrial use case, which is part of the European-funded project AIDOaRt. Our approach contributes to automated decision support to simulation and architecture design through model-driven optimisation while preserving the organisation's engineering practices.

This paper describes the Architecture Life Cycle Effect Analysis (ALCEA) method, a structured method for evaluating proposed new architectures for software-intensive systems. The method evaluates a proposed architecture by quantifying its effect on the performance of system life-cycle phases. The method is instantiated by identifying (i) the relevant life-cycle phases of the system under investigation and (ii) a set of evaluation functions that capture, in terms of basic factors, the effect of different architectural decisions on key life-cycle performance attributes, such as revenue, operating resources, and investments. The method results in a transparent cost and revenue structure, documented in a tabular form, based on quantifiable factors from the developing organization. The results of the method can be used directly as part of a business case, and their robustness can be estimated by sensitivity analysis. The ALCEA method is designed for system-level architectural analysis, covering both software and hardware aspects. In this paper, we introduce the ALCEA method and provide a detailed example of how to apply it in the evolution of embedded systems. Moreover, we share early experiences of using the method in large-scale industrial settings.

This paper focuses on sensor data management in the Solar Car Optimized Route Estimation (SCORE) system which is currently being developed. The data collection is split into two phases: the a priori phase related to the environment and the real time phase related to the solar vehicle. The hardware and software architecture for a priori phase is illustrated in detail, while the real time phase is discussed as a part of the embedded car computer. The full system description for SCORE is presented with guidelines for the future work and implementation of it.