Abstract Smart architectures are increasingly being used in current software development. Smart user interfaces, smart homes, or smart buildings are becoming common examples in the new era of smart cities. Software architectures usually related to these domains need to be adapted and reconfigured at run-time, for example, to provide new services, react to user interaction, or due to changes decided from the business logic of the application. Component-based techniques are a suitable way to carry out this kind of adaptation, as dynamic reconfiguration operations can be applied to the architecture. In this paper, we address run-time generation of component-based applications, taking the abstract definitions of their architecture as a reference, in addition to a set of available components. The process calculates the best configuration of components from the abstract definition by applying a trading approach based on an adapted A* algorithm. This algorithm uses heuristics based on syntactic and semantic information obtained from the component definitions. A case study related to mashup user interfaces formed by coarse-grained components is also explained. In short, the results show the usefulness of heuristics and suitable execution times for building the best configurations.

中文翻译：

---

用于在运行时构建智能架构的基于启发式的中介

摘要 当前的软件开发中越来越多地使用智能架构。智能用户界面、智能家居或智能建筑正在成为智慧城市新时代的常见例子。通常与这些领域相关的软件架构需要在运行时进行调整和重新配置，例如，提供新服务、对用户交互做出反应，或者由于应用程序的业务逻辑决定的更改。基于组件的技术是执行这种适应的合适方式，因为动态重新配置操作可以应用于架构。在本文中，我们讨论了基于组件的应用程序的运行时生成，除了一组可用的组件外，还将其架构的抽象定义作为参考。该过程通过应用基于自适应 A* 算法的交易方法，根据抽象定义计算组件的最佳配置。该算法使用基于从组件定义中获得的句法和语义信息的启发式方法。还解释了与由粗粒度组件形成的混搭用户界面相关的案例研究。简而言之，结果显示了启发式方法和合适的执行时间对于构建最佳配置的有用性。