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.