Research on Parallel Yen Algorithms on GPUs Using CUDA

With the arrival of the intermodality era, to design fast and efficientKshortest paths (KSP) algorithms becomes gradually one of the core technologies in traveler information systems.Yenis a classical algorithm for KSP. However,Yenis time-consuming. In view of powerful general-purpose computing capabilities, GPU(Graphics Processing Units) has received increasing attention in various fields. Based on CUDA software development environment, combined with the structure of theYenalgorithm itself, this paper proposed two parallel algorithms forYen. The first parallel algorithm computes candidate shortest paths for very possible deviation nodes in parallel. The second one computes candidate shortest paths in serial, but computes very candidate path in parallel. Finally, the efficiency of the two parallel algorithms was tested through experiments.

Similar Execution Path Generation Based on Backward Symbolic Analysis

Similar execution paths generation is one of the fundamental tasks in code analysis and detection. The current methods usually target to the program behavior or program structure, and change the value of key predicates, but these methods has a low effectiveness due to the lack of the necessary guidance information, Meanwhile, the predicates set has a large size and usually hard to solve, thus it will reduce the analyze precision as well. A technique of similar execution paths generation based on dynamic synergy bidirectional mapping is proposed in this paper. According to extend the shape of Control Flow Graph and use the backward symbolic analysis, the weakest precondition of the candidate path is generated, which can be used as the guidance information to generate pointed similar execution paths set according to the edit distance via changing the distance factor. The experimental results show that this method has the advantage of precision and anti-inference.

Application of Spatial and Fourier Transforms to the Synthesis of Path Generation Mechanisms

This work presents an approach for the invariant characterization, storage and search of planar path generation mechanisms. Spatial transformations and Fourier descriptors are used to characterize the path of four-bar planar mechanisms. The invariant characteristic information is then stored in a database. The resulting characteristic information for the path is invariant under rigid transformations — translation, rotation and scaling. Solution comparison and matching techniques are discussed and implemented to evaluate the deviation of a desired path to candidate paths stored in a database. A description of each method and the relative performance of file development and search methods are generated for four-bar mechanisms. Each method may be applied to any path generation planar mechanisms since the output path is characterized and stored in a database for solution comparison. The specific type of mechanism that generates the path is not necessarily required for identification of candidate solutions. A database of 7,845 candidate path solutions is developed to evaluate each method. The database of path solutions and related mechanism definitions are based on crank-rocker, four bar mechanism solutions developed by Hrones and Nelson (1951).