SDN simplifies network management by relying on declarativity (high-level interface) and expressiveness (network flexibility). We propose a solution to support those features while preserving high robustness scalability as needed in carrier-grade networks. Our is based (i) two-layer architecture separating connectivity optimization tasks; (ii) centralized optimizer called framework, which translates high-level goals expressed almost natural language into compliant configurations. evaluation...

Several recent publications have studied the use of Mixed Integer Programming (MIP) for finding an optimal decision tree, that is, best tree under formal requirements on accuracy, fairness or interpretability predictive model. These used MIP to deal with hard computational challenge such trees. In this paper, we introduce a new efficient algorithm, DL8.5, trees, based itemset mining techniques. We show approach outperforms earlier approaches several orders magnitude, both numerical and...

SDN simplifies network management by relying on declarativity (high-level interface) and expressiveness (network flexibility). We propose a solution to support those features while preserving high robustness scalability as needed in carrier-grade networks. Our is based (i) two-layer architecture separating connectivity optimization tasks; (ii) centralized optimizer called framework, which translates high-level goals expressed almost natural language into compliant configurations. evaluation...

In this paper, we propose a pragmatic approach to improve reproducibility of experimental analyses traffic engineering (TE) algorithms, whose implementation, evaluation and comparison are currently hard replicate. Our envisioned goal is enable universally-checkable experiments existing future TE algorithms. We describe the design implementation REPETITA, software framework that implements common functions, automates setup, eases comparisons (in terms solution quality, execution time, etc.)...

Segment Routing (SR) is a powerful tool to solve traffic engineering in large networks. It enables steering the along any arbitrary network path while limiting scalability issues as routers do not need maintain global state. Mathematical programming approaches proposed so far for SR either scale well with size of topology or impose strong limit on number possible detours (typically at most one). Moreover they support fully by ignoring adjacency segments. This paper leverages column...

Abstract Background Assessing the impact of government responses to Covid-19 is crucial contain pandemic and improve preparedness for future crises. We investigate here non-pharmaceutical interventions (NPIs) infection threats on daily evolution cross-border movements people during pandemic. use a unique database Facebook users’ mobility, rely regression machine learning models identify role containment policies. Permutation techniques allow us compare predictive power these two categories...

The branch-and-bound algorithm based on decision diagrams is a framework for solving discrete optimization problems with dynamic programming formulation. It works by compiling series of bounded-width that can provide lower and upper bounds any given subproblem. Eventually, every part the search space will be either explored or pruned algorithm, thus proving optimality. This paper presents new ingredients to speed up exploiting structure models. key idea prevent repeated expansion nodes...

In the railway domain, an interlocking is a computerised system that controls signalling objects in order to allow safe operation of train traffic. Each makes use particular data, called application reflects track layout station under control. The verification and validation data are performed manually thus error-prone costly. this paper, we explain how built executable model NuSMV based on data. We also detail tool have developed translate into our automatically. Finally show could verify...

Decision Trees (DTs) are widely used Machine Learning (ML) models with a broad range of applications. The interest in these has increased even further the context Explainable AI (XAI), as decision trees limited depth very interpretable models. However, traditional algorithms for learning DTs heuristic nature; they may produce that suboptimal quality under constraints. We introduce PyDL8.5, Python library to infer depth-constrained Optimal (ODTs). PyDL8.5 provides an interface DL8.5,...

In the railway domain, an interlocking is system ensuring safe train traffic inside a station by controlling its active elements such as signals or points. Modern interlockings are configured using particular data, called application reflecting track layout and defining actions that can take. The safety of relies thereby on data correctness, errors them cause issues derailments collisions. Given high level required system, verification critical concern. addition to safety, must also ensure...

Multi-Valued Decision Diagrams (MDDs) are instrumental in modeling combinatorial problems with Constraint Programming.In this paper, we propose a related data structure called sMDD (semi-MDD) where the central layer of diagrams is non-deterministic.We show that it easy and efficient to transform any table (set tuples) into an sMDD.We also introduce new filtering algorithm, Compact-MDD, which based on bitwise operations, can be applied both MDDs sMDDs.Our experimental results practical...