Abstract

Abstract

Abstract
Strategic term rewriting and attribute grammars are two powerful programming techniques widely used in language engineering. The former relies on strategies (recursion schemes) to apply term rewrite rules in defining transformations, while the latter is suitable for expressing context-dependent language processing algorithms. Each of these techniques, however, is usually implemented by its own powerful and large processor system. As a result, it makes such systems harder to extend and to combine. We present the embedding of both strategic tree rewriting and attribute grammars in a zipper-based, purely functional setting. The embedding of the two techniques in the same setting has several advantages: First, we easily combine/zip attribute grammars and strategies, thus providing language engineers the best of the two worlds. Second, the combined embedding is easier to maintain and extend since it is written in a concise and uniform setting. We show the expressive power of our library in optimizing Haskell let expressions, expressing several Haskell refactorings and solving several language processing tasks for an Oberon-0 compiler. © 2022, Springer Nature Switzerland AG.

Abstract
First-mile logistics tackles the movement of products from retailers to a warehouse or distri-bution centre. This first step towards the end customer has been pushed by large e-commerce platforms forming extensive networks of partners and is critical for fast deliveries. First-mile pickup requires efficient methods different from those developed for last-mile delivery, among other reasons due to the complexity of cargo features and volume - increasing the relevance of advanced packing methods. More importantly, the problem is essentially dynamic and the pickup process, in which the vehicle is initially empty, is much more flexible to react to disruptions arising when the vehicles are en route. We model the static first-mile pickup problem as a vehicle routing problem for a hetero-geneous fleet, with time windows and three-dimensional packing constraints. Moreover, we propose an approach to tackle the dynamic problem, in which the routes can be modified to accommodate disruptions - new customers' demands and modified requests of known customers that are arriving while the initially established routes are being covered. We propose three reactive strategies for addressing the disruptions depending on the number of vehicles available, and study their results on a newly generated benchmark for dynamic problems. The results allow quantifying the impact of disruptions depending on the strategy used and can help the logistics companies to define their own strategy, considering the characteristics of their customers and products and the available fleet.

Abstract
Decision trees are often preferred when implementing Machine Learning in embedded systems for their simplicity and scalability. Hoeffding Trees are a type of Decision Trees that take advantage of the Hoeffding Bound to allow them to learn patterns in data without having to continuously store the data samples for future reprocessing. This makes them especially suitable for deployment on embedded devices. In this work we highlight the features of a HLS implementation of the Hoeffding Tree. The implementation parameters include the feature size of the samples (D), the number of output classes (K), and the maximum number of nodes to which the tree is allowed to grow (Nd). We target a Xilinx MPSoC ZCU102, and evaluate: the design's resource requirements and clock frequency for different numbers of classes and feature size, the execution time on several synthetic datasets of varying sizes (N) and the execution time and accuracy for two datasets from UCI. For a problem size of D=3, K=5, and N=40000, a single decision tree operating at 103MHz is capable of 8.3x faster inference than the 1.2 GHz ARM Cortex-A53 core. Compared to a reference implementation of the Hoeffding tree, we achieve comparable classification accuracy for the UCI datasets.

Abstract