Semi-supervised model adaptation for statistical machine translation

Statistical machine translation systems are usually trained on large amounts of bilingual text (used to learn a translation model), and also large amounts of monolingual text in the target language (used to train a language model). In this article we explore the use of semi-supervised model adaptation methods for the effective use of monolingual data from the source language in order to improve translation quality. We propose several algorithms with this aim, and present the strengths and weaknesses of each one. We present detailed experimental evaluations on the French–English EuroParl data set and on data from the NIST Chinese–English large-data track. We show a significant improvement in translation quality on both tasks.     

Optimal job sequence determination and operation machine allocation in flexible manufacturing systems: an approach using adaptive hierarchical ant colony algorithm

In this paper, an Adaptive Hierarchical Ant Colony Optimization (AHACO) has been proposed to resolve the traditional machine loading problem in Flexible Manufacturing Systems (FMS). Machine loading is one of the most important issues that is interlinked with the efficiency and utilization of FMS. The machine loading problem is formulated in order to minimize the system unbalance and maximize the throughput, considering the job sequencing, optional machines and technological constraints. The performance of proposed AHACO has been tested over a number of benchmark problems taken from the literature. Computational results indicate that the proposed algorithm is more effective and produces promising results as compared to the existing solution methodologies in the literature. The evaluation and comparison of system efficiency and system utilization justifies the supremacy of the algorithm. Further, results obtained from the proposed algorithm have been compared with well known random search algorithm viz. genetic algorithm, simulated annealing, artificial Immune system, simple ant colony optimization, tabu search etc. In addition, the algorithm has been tested over a randomly generated problem set of varying complexities; the results validate the robustness and scalability of the algorithm utilizing the concepts of ‘heuristic gap’ and ANOVA analysis.     

Attack Resilient and Efficient Protocol based on Greedy Perimeter Coordinator Routing—Mobility Awareness for Preventing the Attack in the VANET

Wireless Personal Communications - A Greedy Perimeter Coordinator Routing and Mobility Awareness (GPCR-MA) vehicular routing is a widely accepted routing protocol for VANET (Vehicular Ad hoc...     

Numerical investigation of heat transfer and friction factor in ribbed triangular duct solar air heater using Computational fluid dynamics (CFD)

Journal of Mechanical Science and Technology - The Computational fluid dynamics (CFD) based analysis is carried out to investigate the thermal and hydraulic performance of circular rib roughened...     

A mathematical model and solution methods for rail freight transportation planning in an Indian food grain supply chain

This paper addresses the rail transportation of food grains undertaken by Food Corporation of India (FCI) to meet the requirements of the food security programme called Public Distribution System (PDS). The research focuses on improving the allocation of railway rakes transporting food grains to a set of storage warehouses. A penalty factor based approach is adopted to represent the considerations in transportation planning and three penalty factors such as rake penalty factor, weekly penalty factor and capacity utilization penalty factor are introduced for the purpose. The single source - multiple destination problem is formulated and solved using exact method to minimize the sum of these three penalty factor values, termed total penalty. Further, a heuristic named optimum rake allocation algorithm is developed and tested using a set of 35 problem instances. The proposed heuristic is found to be highly efficient in terms of solution quality and computation time. A case study of FCI Kerala Region is also carried out to validate the formulated model and the proposed heuristic. The work provides valuable insights into the practical issues encountered in rail freight transportation planning and proposes an effective solution methodology to address them.     

Lattice Boltzmann Simulation of Two- and Three-Dimensional Incompressible Thermal Flows

This work is concerned with the application of the thermal lattice Boltzmann method (TLBM) to compute incompressible two- and three-dimensional flows in cavities. Two convection test cases, namely, the laminar flow in a differentially heated square cavity and a cubic cavity, are numerically analyzed through TLBM. The internal energy density distribution function approach with two three-dimensional particle velocity models, namely, the 15-velocity and the 19-velocity, and a two-dimensional model, namely, the nine-velocity, have been used in the present work. Computations are carried out for laminar flows in a differentially heated square cavity and a cubical cavity (Rayleigh numbers = 10³ to 10⁵). The boundary conditions used are stable and of good accuracy. To lend credibility to the thermal lattice Boltzmann model square cavity results, they are further compared with those obtained from a finite-difference-based code developed for this purpose.     

Loading Rate Effect on Nanohardness of Soda-Lime-Silica Glass

To understand how hardness, the key design parameter for applications of brittle solids such as glass concerning contact deformation, is affected by loading rate variation, nanoindentation with a Berkovich tip was used to measure the nanohardness of a 330-μm-thick soda-lime-silica glass as a function of loading rate (1 to 1000 mN·s⁻¹). The results showed for the very first time that, with variations in the loading rate, there was a 6 to 9 pct increase in the nanohardness of glass up to a threshold loading rate (TLR), whereafter it did not appreciably increase with further increase in loading rate. Further, the nanohardness data showed an indentation size effect (ISE) that obeyed the Meyer’s law. These observations were explained in terms of a strong shear stress component developed just beneath the nanoindenter and the related shear-induced deformation processes at local microstructural scale weak links. The significant or insignificant presence of shear-induced serrations in load depth plots and corresponding scanning electron microscopic evidence of a strong or mild presence of shear deformation bands in and around the nanoindentation cavity supported such a rationalization. Finally, a qualitative picture was developed for different deformation processes induced at various loading rates in glass.