Genetic Threading

The biological function of proteins is dependent, to a large extent, on their native three dimensional conformation. Thus, it is important to know the structure of as many proteins as possible. Since experimental methods for structure determination are very tedious, there is a significant effort to calculate the structure of a protein from its linear sequence. Direct methods of calculating structure from sequence are not available yet. Thus, an indirect approach to predict the conformation of protein, called threading, is discussed. In this approach, known structures are used as constraints, to restrict the search for the native conformation. Threading requires finding good alignments between a sequence and a structure, which is a major computational challenge and a practical bottleneck in applying threading procedures. The Genetic Algorithm paradigm, an efficient search method that is based on evolutionary ideas, is used to perform sequence to structure alignments. A proper representation is discussed in which genetic operators can be effectively implemented. The algorithm performance is tested for a set of six sequence/structure pairs. The effects of changing operators and parameters are explored and analyzed.     

Resource-efficient routing and scheduling of time-constrained streaming communication on networks-on-chip

Network-on-chip-based multiprocessor systems-on-chip are considered as future embedded systems platforms. One of the steps in mapping an application onto such a parallel platform involves scheduling the communication on the network-on-chip. This paper presents different scheduling strategies that minimize resource usage by exploiting all scheduling freedom offered by networks-on-chip. It also introduces a technique to take the dynamism in applications into account when scheduling the communication of an application on the network-on-chip while minimizing the resource usage. Our experiments show that resource-utilization is improved when compared to existing techniques.     

Simple fiber-optic confocal microscopy with nanoscale depth resolution beyond the diffraction barrier

A novel fiber-optic confocal approach for ultrahigh depth-resolution (相似文献   

Learning to transform time series with a few examples

We describe a semi-supervised regression algorithm that learns to transform one time series into another time series given examples of the transformation. This algorithm is applied to tracking, where a time series of observations from sensors is transformed to a time series describing the pose of a target. Instead of defining and implementing such transformations for each tracking task separately, our algorithm learns a memoryless transformation of time series from a few example input-output mappings. The algorithm searches for a smooth function that fits the training examples and, when applied to the input time series, produces a time series that evolves according to assumed dynamics. The learning procedure is fast and lends itself to a closed-form solution. It is closely related to nonlinear system identification and manifold learning techniques. We demonstrate our algorithm on the tasks of tracking RFID tags from signal strength measurements, recovering the pose of rigid objects, deformable bodies, and articulated bodies from video sequences. For these tasks, this algorithm requires significantly fewer examples compared to fully-supervised regression algorithms or semi-supervised learning algorithms that do not take the dynamics of the output time series into account.     

Efficient Regular Data Structures and Algorithms for Dilation, Location, and Proximity Problems

In this paper we investigate data structures obtained by a recursive partitioning of the multi- dimensional input domain into regions of equal size. One of the best known examples of such a structure is the quadtree . It is used here as a basis for more complex data structures. We also provide multidimensional versions of the stratified tree by van Emde Boas [vEB]. We show that under the assumption that the input points have limited precision (i.e., are drawn from the integer grid of size u ) these data structures yield efficient solutions to many important problems. In particular, they allow us to achieve O(log log u) time per operation for dynamic approximate nearest neighbor (under insertions and deletions) and exact on-line closest pair (under insertions only) in any constant number of dimensions. They allow O(log log u) point location in a given planar shape or in its expansion (dilation by a ball of a given radius). Finally, we provide a linear time (optimal) algorithm for computing the expansion of a shape represented by a region quadtree. This result shows that the spatial order imposed by this regular data structure is sufficient to optimize the operation of dilation by a ball. Received January 19, 1999; revised November 4, 1999.     

Efficient matching of nonrectangular shapes

Most known two-dimensional matching algorithms have a rectangular text (image) and rectangular pattern (template) as their input. These matching algorithms perform a row by row analysis followed by some column processing. These techniques are only efficient if all the rows are of equal length, hence the necessity for rectangular patterns.We present a novel method for analysing patterns with rows of different lengths. This enables finding all occurrences of a nonrectangular figure of heightm in ann×n text in time . We make use of thesmaller matching problem.The smaller matching problem is that of finding all appearances of a numerical one dimensional pattern in a numerical one-dimensional text, where every element of the pattern is no greater than the corresponding text element.Partially supported by NSF grant CCR-88-03641 and a University of Maryland full year research award.Supported by a University of Maryland Graduate Fellowship, an ACM Samuel M. Alexander Fellowship and NSF grant CCR-88-03641.     

Numerical Simulation of Submarine Groundwater Flow in the Coastal Aquifer at the Palmahim Area, the Mediterranean Coast of Israel

The lower sub-aquifers of the Mediterranean coastal aquifer of Israel, at the Palmahim area, are hypothesized to be laterally blocked to connection with the sea, and thus to seawater intrusion. This is mostly due to the detection of fresh water bodies at these sub-aquifers. This study examine this hypothesis by using two dimensional numerical model simulations of the groundwater flow system at this area, which conducted in order to reveal which hydrogeological setting enables the existence of these fresh water bodies in the lower sub-aquifers and to assess the on-land pumping rates that will prevent their salinization. The hydrogeological settings were examined by steady state simulations followed by simulations of the last 15,000 years sea level changes. These simulations imply that the presence of fresh water in the lower sub-aquifer, whether blocked or connected to the sea, requires offshore separation between the upper and lower sub-aquifers. On-land pumping simulations, with a well located inside the lower sub-aquifers at the shoreline, show a maximum pumping rate of 250 m³/m strip width/year, hereafter m²/year, to prevent the salinization of the lower sub-aquifers. The various pumping scenarios revealed differences in salinization trends between the scenarios with impermeable separating layers and those with semi permeable layers. Scenarios with extreme pumping rates emphasize these differences, and together with field test, can allow assessing the amount of separation between the sub-aquifers.     

Evaluation of suspended load transport rate using transport formulas and artificial neural network models （Case study： Chelchay Catchment）＊

Accurate estimation of sediment load or transport rate is very important to a wide range of water resources projects. This study was undertaken to determine the most appropriate model to predict suspended load in the Chelchay Watershed, northeast of Iran. In total, 59 data series were collected from four gravel bed-rivers and a sand bed river and two depth integrating suspended load samplers to evaluate nine suspended load formulas and feed forward backpropagation Artificial Neural Network (ANN) structures. Although the Chang formula with higher correlation coefficient (r = 0.69) and lower Root Mean Square Error (RMSE = 0.013) is the best suspended load predictor among the nine studied formulas, the ANN models significantly outperform traditional suspended load formulas and show their superior performance for all statistical parameters. Among different ANN structures two models including 4 inputs, 4 hidden and one output neurons, and 4 inputs, 4 and one hidden and one output neurons provide the best simulation with the RMSE values of 0.0009 and 0.001, respectively.     

Efficient biodegradation of naphthalene by a newly characterized indigenous Achromobacter sp. FBHYA2 isolated from Tehran Oil Refinery Complex

A bacterial strain, FBHYA2, capable of degrading naphthalene, was isolated from the American Petroleum Institute (API) separator of the Tehran Oil Refinery Complex (TORC). Strain FBHYA2 was identified as Achromobacter sp. based on physiological and biochemical characteristics and also phylogenetic similarity of 16S rRNA gene sequence. The optimal growth conditions for strain FBHYA2 were pH 6.0, 30 °C and 1.0% NaCl. Strain FBHYA2 can utilize naphthalene as the sole source of carbon and energy and was able to degrade naphthalene aerobically very fast, 48 h for 96% removal at 500 mg/L concentration. The physiological response of Achromobacter sp., FBHYA2 to several hydrophobic chemicals (aliphatic and aromatic hydrocarbons) was also investigated. No biosurfactant was detected during bacterial growth on any aliphatic/aromatic hydrocarbons. The results of hydrophobicity measurements showed no significant difference between naphthalene- and LB-grown cells. The capability of the strain FBHYA2 to degrade naphthalene completely and rapidly without the need to secrete biosurfactant may make it an ideal candidate to remediate polycyclic aromatic hydrocarbon (PAH)-contaminated sites.     

Optimum loading of machines in a flexible manufacturing system using a mixed-integer linear mathematical programming model and genetic algorithm

Machine loading problem in a flexible manufacturing system (FMS) encompasses various types of flexibility aspects pertaining to part selection and operation assignments. The evolution of flexible manufacturing systems offers great potential for increasing flexibility by ensuring both cost-effectiveness and customized manufacturing at the same time. This paper proposes a linear mathematical programming model with both continuous and zero-one variables for job selection and operation allocation problems in an FMS to maximize profitability and utilization of system. The proposed model assigns operations to different machines considering capacity of machines, batch-sizes, processing time of operations, machine costs, tool requirements, and capacity of tool magazine. A genetic algorithm (GA) is then proposed to solve the formulated problem. Performance of the proposed GA is evaluated based on some benchmark problems adopted from the literature. A statistical test is conducted which implies that the proposed algorithm is robust in finding near-optimal solutions. Comparison of the results with those published in the literature indicates supremacy of the solutions obtained by the proposed algorithm for attempted model.