Technical reports

Abstract. Numerous studies continue to explore the potential of social interactions between people in Free/Libre Open Source Software (FLOSS) environments. While the dynamics of interactions in these environments can be understood from different perspectives, we put a particular focus on any interactions resulting in knowledge transfer and acquisition. As learning platforms, FLOSS communities provide immense opportunities for improving software engineering skills. People who engage in FLOSS activities both acquire and improve their software development skills. For this reason, it is very helpful to understand how these learning interactions occur. In this paper, we make use of the decision miner in process mining to conduct our analysis. The purpose of such an endeavour is twofold. Firstly, we provide empirical insights into how people learn while exchanging emails in FLOSS mailing archives. Lastly, we go a step further by providing insights behind the motivation into learning participants' decisions on their learning paths.

Fog computing aims at extending the Cloud by bringing computational power, storage and communication capabilities to the edge of the network, in support of the IoT. Segmentation, distribution and adaptive deployment of functionalities over the continuum from Things to Cloud are challenging tasks, due to the intrinsic heterogeneity, hierarchical structure and very large scale infrastructure they will have to exploit. In this paper we propose a simple, yet general, model to support the QoS-aware deployment of multi-component IoT applications over Fog infrastructures. The model describes operational systemic qualities of the available infrastructure (latency and bandwidth), interactions among software components and Things, and business policies. Algorithms to determine eligible deployment plans for an application over a Fog infrastructure are presented. A Java tool, FogTorch, based on the proposed model has been prototyped.

Process mining is a relatively new field that encompasses powerful data and process analytics techniques for understanding processes from event data. In addition to these main techniques, it provides means enabling a pictorial representation of the occurrence of events over time. By applying such visualizations to event data from Free/Libre Open Source Software (FLOSS) environments, we get a complete understanding of how certain activities take place within such environments over time. Particularly, given the increasing interests in learning paradigms present in FLOSS communities, we believe that a temporal visual representation of learning events can yield great benefits. In this paper, we make use of the dotted chart in process mining to model and present a representation of learning behaviours over time for FLOSS participants

FLOSS environments have been proved to provide an interesting learning platform for software engineers. Research suggests that people partaking in both technical and non-technical activities in FLOSS prjects are more likely to positively improve their software engineering skills. To this end, there are propositions to involve computer science and software engineering students in formal higher institutions of learning, in participating in FLOSS projects in order to give them an opportunity to develop their programming capacity by working on real-life projects. While some empirical studies have been conducted to provide some lights on learning processes in FLOSS environments, there is limited or no work done pertaining to understanding social structures during this process of knowledge transfer and acquisition. In this paper, we make use of social network analysis techniques in order to provide insights related to the emerging of social structures from FLOSS repositories from an educational point of view. We hope that these educational structures will enhance both the understanding with regards to how learning occurs in these communities and especially, the frequency of participants' involvement that culminates into learning.

We propose an improvement of the Approximated Projected Perspective Reformulation (AP^2R) of [Frangioni, Furini, Gentile, Computational Optimization and Applications, 2016] for the case in which constraints linking the binary variables exist. The new approach requires to solve the Perspective Reformulation (PR) once, and then use the corresponding dual information to reformulate the problem prior to applying AP^2R, thereby combining the root bound quality of the PR with the reduced relaxation computing time of AP^$R. Computational results for the cardinality-constrained Mean-Variance portfolio optimization problem show that the new approach is competitive with state-of-the-art ones.

We propose a family of proximal bundle methods for minimizing sum-structured convex nondifferentiable functions which require two slightly uncommon assumptions, that are satisfied in many relevant applications: Lipschitz continuity of the functions and oracles which also produce upper estimates on the function values. In exchange, the methods: i) use upper models of the functions that allow to estimate function values at points where the oracle has not been called; ii) provide the oracles with more information about when the function computation can be interrupted, possibly diminishing their cost; iii) allow to skip oracle calls entirely for some of the component functions, not only at ``null steps'' but also at ``serious steps''; iv) provide explicit and reliable a-posteriori estimates of the quality of the obtained solutions; v) work with all possible combinations of different assumptions on the oracles. We also discuss introduction of constraints (or, more generally, of easy components) and use of (partly) aggregated models.