Matrix-vector multiplication and triangular linear solver using GPGPU for symmetric positive definite matrices derived from elliptic equations

The modern GPUs are well suited for intensive computational tasks and massive parallel computation. Sparse matrix multiplication and linear triangular solver are the most important and heavily used kernels in scientific computation, and several challenges in developing a high performance kernel with the two modules is investigated. The main interest it to solve linear systems derived from the elliptic equations with triangular elements. The resulting linear system has a symmetric positive definite matrix. The sparse matrix is stored in the compressed sparse row (CSR) format. It is proposed a CUDA algorithm to execute the matrix vector multiplication using directly the CSR format. A dependence tree algorithm is used to determine which variables the linear triangular solver can determine in parallel. To increase the number of the parallel threads, a coloring graph algorithm is implemented to reorder the mesh numbering in a pre-processing phase. The proposed method is compared with parallel and serial available libraries. The results show that the proposed method improves the computation cost of the matrix vector multiplication. The pre-processing associated with the triangular solver needs to be executed just once in the proposed method. The conjugate gradient method was implemented and showed similar convergence rate for all the compared methods. The proposed method showed significant smaller execution time.

Three time-based scale formulations for the two-stage lot sizing and scheduling in process industries

In this paper, we propose three novel mathematical models for the two-stage lot-sizing and scheduling problems present in many process industries. The problem shares a continuous or quasi-continuous production feature upstream and a discrete manufacturing feature downstream, which must be synchronized. Different time-based scale representations are discussed. The first formulation encompasses a discrete-time representation. The second one is a hybrid continuous-discrete model. The last formulation is based on a continuous-time model representation. Computational tests with state-of-the-art MIP solver show that the discrete-time representation provides better feasible solutions in short running time. On the other hand, the hybrid model achieves better solutions for longer computational times and was able to prove optimality more often. The continuous-type model is the most flexible of the three for incorporating additional operational requirements, at a cost of having the worst computational performance. Journal of the Operational Research Society (2012) 63, 1613-1630. doi:10.1057/jors.2011.159 published online 7 March 2012

Photodynamic Efficiency of Cationic meso-Porphyrins at Lipid Bilayers: Insights from Molecular Dynamics Simulations

Porphyrin derivatives have applications as photoactive drugs in photodynamic therapy. However, little is known about their interactions with phospholipid membranes at the molecular level. We employed molecular dynamics simulations to model the binding between a series of cationic meso-(N-methyl-4-pyridinium)phenylporphyrins and anionic phosphatidylglycerol lipid bilayers. This was done in the presence of molecular oxygen within the membrane. The ability of various porphyrins to cause photodamage was quantified in terms of their immersion depth and degree of exposition to a higher oxygen concentration inside the membrane. Simulations showed that the photodynamic efficiency could be improved as the number of hydrophobic phenyl substituents attached to the porphyrinic ring increased. In the specific case of porphyrins containing two hydrophobic and two charged substituents, the cis isomer was significantly more efficient than the trans. These results correlate well with previous experimental observations. They highlight the importance of both the total charge and amphiphilicity of the photosensitizer for its performance in photodynamic therapy.

Augmented Lagrangian method with nonmonotone penalty parameters for constrained optimization

At each outer iteration of standard Augmented Lagrangian methods one tries to solve a box-constrained optimization problem with some prescribed tolerance. In the continuous world, using exact arithmetic, this subproblem is always solvable. Therefore, the possibility of finishing the subproblem resolution without satisfying the theoretical stopping conditions is not contemplated in usual convergence theories. However, in practice, one might not be able to solve the subproblem up to the required precision. This may be due to different reasons. One of them is that the presence of an excessively large penalty parameter could impair the performance of the box-constraint optimization solver. In this paper a practical strategy for decreasing the penalty parameter in situations like the one mentioned above is proposed. More generally, the different decisions that may be taken when, in practice, one is not able to solve the Augmented Lagrangian subproblem will be discussed. As a result, an improved Augmented Lagrangian method is presented, which takes into account numerical difficulties in a satisfactory way, preserving suitable convergence theory. Numerical experiments are presented involving all the CUTEr collection test problems.

PERFORMANCE EVALUATION OF RESOURCE-AWARE BUSINESS PROCESSES USING STOCHASTIC AUTOMATA NETWORKS

In this work, we study the performance evaluation of resource-aware business process models. We define a new framework that allows the generation of analytical models for performance evaluation from business process models annotated with resource management information. This framework is composed of a new notation that allows the specification of resource management constraints and a method to convert a business process specification and its resource constraints into Stochastic Automata Networks (SANs). We show that the analysis of the generated SAN model provides several performance indices, such as average throughput of the system, average waiting time, average queues size, and utilization rate of resources. Using the BP2SAN tool - our implementation of the proposed framework - and a SAN solver (such as the PEPS tool) we show through a simple use-case how a business specialist with no skills in stochastic modeling can easily obtain performance indices that, in turn, can help to identify bottlenecks on the model, to perform workload characterization, to define the provisioning of resources, and to study other performance related aspects of the business process.