Abstract In this paper we suggest a simple algorithm for merging the traffic of a main road with on-ramp traffic for collaborative Connected and Autonomous Vehicles (CAVs). The main asset of the algorithm being that most values can be formally computed, we quantitatively express the global impact of the merging strategy on the input and output flows of the merging area. For instance, how platoons emerge, their lengths, the space between them and their traffic patterns. We refer to the literature for all local aspects of synchronisation of cars. Here we simply assume that cars can execute the merging algorithm which requires that the cars are aware of a sequence number and operates following the sequence. Our analysis focuses on saturated flows and the control of collective behavior.

MRI, a widespread non-invasive medical imaging modality, is highly sensitive to patient motion. Despite many attempts over the years, motion correction remains a difficult problem and there is no general method applicable to all situations. We propose a retrospective method for motion estimation and correction to tackle the problem of in-plane rigid-body motion, apt for classical 2D Spin-Echo scans of the brain, which are regularly used in clinical practice. Due to the sequential acquisition of k-space, motion artifacts are well localized. The method leverages the power of deep neural networks to estimate motion parameters in k-space and uses a model-based approach to restore degraded images to avoid "hallucinations". Notable advantages are its ability to estimate motion occurring in high spatial frequencies without the need of a motion-free reference. The proposed method operates on the whole k-space dynamic range and is moderately affected by the lower SNR of higher harmonics. As a proof of concept, we provide models trained using supervised learning on 600k motion simulations based on motion-free scans of 43 different subjects. Generalization performance was tested with simulations as well as in-vivo. Qualitative and quantitative evaluations are presented for motion parameter estimations and image reconstruction. Experimental results show that our approach is able to obtain good generalization performance on simulated data and in-vivo acquisitions. We provide a Python implementation at https://gitlab.unige.ch/Oscar.Dabrowski/sismik_mri/.

Abstract We consider the car following problem for a set of autonomous vehicles following each other on either an infinite or circular road. The behavior of each car is specified by its "speed regulator", a device that decides to increase or decrease the speed of the car as a function of the head-tail distance to its predecessor and the speed of both cars. A collective behavior emerges that corresponds to previously proposed cellular automata traffic models. We further analyze the traffic patterns of the system in the long term, as governed by the speed regulator and we study under which conditions traffic patterns of maximum flow can or cannot be reach. We show the existence of suboptimal flow conditions that require external coordination mechanisms (that we do not consider in this paper) in order to reach the optimal flow achievable with the given density. In contrast with other approaches, we do not try to reproduce observed or measured traffic patterns. We analyze a deterministic speed regulator in order to decipher the emergent dynamics, and to ponder what maneuvers can be safely performed. Here, we restrict our attention to the car following problem. By comparing our speed regulator with classical models, auch as the Nagel–Schreckenberg and KKW models, we observe that although our regulator is formulated in simple terms, its dynamics share similarities with these models. In particular, the KKW model is designed to reproduce the observed behavior that a trailing car in the synchronization range of the leading car tends to regulate its speed to maintain a constant distance. this same behavior is adopted by our speed regulator, showing that this is a safe way of driving.

Parallel iterative applications often suffer from load imbalance, one of the most critical performance degradation factors. Hence, load balancing techniques are used to distribute the workload evenly to maximize performance. A key challenge is to know when to use load balancing techniques. In general, this is done through load balancing criteria, which trigger load balancing based on runtime application data and/or user-defined information. In the first part of this paper, we introduce a novel, automatic load balancing criterion derived from a simple mathematical model. In the second part, we propose a branch-and-bound algorithm to find the load balancing iterations that lead to the optimal application performance. This algorithm finds the optimal load balancing scenario in polynomial time while, to the best of our knowledge, it has never been addressed in less than an exponential time. Finally, we compare the performance of the scenarios produced by state-of-the-art load balancing criteria relative to the optimal load balancing scenario in synthetic benchmarks and parallel N-body simulations. In the synthetic benchmarks, we observe that the proposed criterion outperforms the other automatic criteria. In the numerical experiments, we show that our new criterion is, on average, 4.9% faster than state-of-the-art load balancing criteria and can outperform them by up to 17.6%. Moreover, we see in the numerical study that the state-of-the-art automatic criteria are at worst 26.43% slower than the optimum and at best 10% slower.

Cardio/cerebrovascular diseases (CVD) have become one of the major health issue in our societies. But recent studies show that the present pathology tests to detect CVD are ineffectual as they do not consider different stages of platelet activation or the molecular dynamics involved in platelet interactions and are incapable to consider inter-individual variability. Here we propose a stochastic platelet deposition model and an inferential scheme to estimate the biologically meaningful model parameters using approximate Bayesian computation with a summary statistic that maximally discriminates between different types of patients. Inferred parameters from data collected on healthy volunteers and different patient types help us to identify specific biological parameters and hence biological reasoning behind the dysfunction for each type of patients. This work opens up an unprecedented opportunity of personalized pathology test for CVD detection and medical treatment.

In parallel iterative applications, computational efficiency is essential for addressing large problems. Load imbalance is one of the major performance degradation factors of parallel applications. Therefore, distributing, cleverly, and as evenly as possible, the workload among processing elements (PE) maximizes application performance. So far, the standard load balancing method consists in distributing the workload evenly between PEs and, when load imbalance appears, redistributing the extra load from overloaded PEs to underloaded PEs. However, this does not anticipate the load imbalance growth that may continue during the next iterations. In this paper, we present a first step toward a novel philosophy of load balancing that unloads the PEs that will be overloaded in the near future to let the application rebalance itself via its own dynamics. Herein, we present a formal definition of our new approach using a simple mathematical model and discuss its advantages compared to the standard load balancing method. In addition to the theoretical study, we apply our method to an application that reproduces the computation of a fluid model with non-uniform erosion. The performance validates the benefit of anticipating load imbalance. We observed up to 16% performance improvement compared to the standard load balancing method.

Since volcanic eruptions are events that can greatly endanger human lives and disrupt human activities such as air traffic, it is of major interest to simulate such phenomena. Volcanic ash transport and dispersion models typically describe particle motion in a turbulent velocity field. Volcanic particles (known as tephra) are advected inside this field from the moment they leave the vent of the volcano until they deposit on the ground. We developed a flexible simulation tool called TETRAS (TEphra TRAnsport Simulator) based on a hybrid Eulerian-Lagrangian model. As this kind of model needs computationally intensive simulations, a parallelization on a distributed memory architecture was developed. Volcanic eruptions involve phenomena occurring at multiple temporal and spatial scales, such as fast rising of particles inside the volcanic column, transport of fine particles over very long range or aggregation of fine particles into larger ones. Thus, we developed a multiscale implementation of the software using MMSF (Multiscale Modeling and Simulation Framework) techniques and tools. This implementation combines short and long-range transport of particles as well as particle aggregation. In this paper, we show the multiscale aspect of the design and implementation. Then, we tackle two load balancing problems that arise from this model: first at the transport code level, where unequal distribution of particles inside the domain directly leads to different load for processors if a naive work distribution is chosen; second at the multiscale level, where the resource allocation for each submodel must be chosen wisely to achieve optimal performances.

Motivation: Bayesian inference is widely used nowadays and relies largely on Markov chain Monte Carlo (MCMC) methods. Evolutionary biology has greatly benefited from the developments of MCMC methods, but the design of more complex and realistic models and the ever growing availability of novel data is pushing the limits of the current use of these methods. Results: We present a parallel Metropolis-Hastings (M-H) framework built with a novel combination of enhancements aimed towards parameter-rich and complex models. We show on a parameter-rich macroevolutionary model increases of the sampling speed up to 35 times with 32 processors when compared to a sequential M-H process. More importantly, our framework achieves up to a twentyfold faster convergence to estimate the posterior probability of phylogenetic trees using 32 processors when compared to the well-known software MrBayes for Bayesian inference of phylogenetic trees. Availability and Implementation: https://bitbucket.org/XavMeyer/hogan. Contact: [email protected]. Supplementary information: Supplementary data are available at Bioinformatics online.

MOTIVATION: Post-translational modifications (PTMs) are important steps in the maturation of proteins. Several models exist to predict specific PTMs, from manually detected patterns to machine learning methods. On one hand, the manual detection of patterns does not provide the most efficient classifiers and requires an important workload, and on the other hand, models built by machine learning methods are hard to interpret and do not increase biological knowledge. Therefore, we developed a novel method based on patterns discovery and decision trees to predict PTMs. The proposed algorithm builds a decision tree, by coupling the C4.5 algorithm with genetic algorithms, producing high-performance white box classifiers. Our method was tested on the initiator methionine cleavage (IMC) and N(α)-terminal acetylation (N-Ac), two of the most common PTMs. RESULTS: The resulting classifiers perform well when compared with existing models. On a set of eukaryotic proteins, they display a cross-validated Matthews correlation coefficient of 0.83 (IMC) and 0.65 (N-Ac). When used to predict potential substrates of N-terminal acetyltransferaseB and N-terminal acetyltransferaseC, our classifiers display better performance than the state of the art. Moreover, we present an analysis of the model predicting IMC for Homo sapiens proteins and demonstrate that we are able to extract experimentally known facts without prior knowledge. Those results validate the fact that our method produces white box models. AVAILABILITY AND IMPLEMENTATION: Predictors for IMC and N-Ac and all datasets are freely available at http://terminus.unige.ch/.

The Branch-and-Bound (B&B) method is a well-known optimization algorithm for solving integer linear programming (ILP) models in the field of operations research. It is part of software often employed by businesses for finding solutions to problems such as airline scheduling problems. It operates according to a divide-and-conquer principle by building a tree-like structure with nodes that represent linear programming (LP) problems. A LP solver commonly used to process the nodes is the simplex method. Nowadays its sequential implementation can be found in almost all commercial ILP solvers. In this paper, we present a hybrid CPU-GPU implementation of the B&B algorithm. The B&B tree is managed by the CPU, while the revised simplex method is mainly a GPU implementation, relying on the CUDA technology of NVIDIA. The CPU manages concurrently multiple instances of the LP solver. The principal difference with a sequential implementation of the B&B algorithm pertains to the LP solver, provided that the B&B tree is managed with the same strategy. We thus compared our GPU-based implementation of the revised simplex to a well-known open-source sequential solver, named CLP, of the COIN-OR project. For given problem densities, we measured a size threshhold beyond which our GPU implementation outperformed its sequential counterpart.

SystemC has become a very popular standardized language for the modeling of System-On-Chip (SoC) devices. However, due to the ever increasing complexity of SoC designs, the ever longer simulation times affect SoC exploration potential and time-to-market. In order to reduce these times, we have developed a parallel SystemC kernel. Because the SystemC semantics require a high level of synchronization which can dramatically affect the performance gains, we investigate in this paper some ways to reduce the synchronization overheads. We validate then our approaches against an academic design model and a real, industrial application.

Future and current radio communications systems based on dynamic channel allocation or CDMA do not require any or very limited frequency planning. This leads to emphasize the importance of coverage predictions in radio planning tools for the future. Current radio propagation predictions are usually based on computations which must be performed separately for each cell in a radio network. Thus, the computation time is directly proportional to the number of cells considered. A scheme named ParFlow can inherently provide coverage predictions for one or several cells without any major effect on the computation time. The capability of the multiple cells simulations of the ParFlow scheme is demonstrated and analyzed for both the computation of composite coverage and the prediction of best server maps. The performance of parallel computations over a network of PC-based workstations is also investigated.