We study the problem of index selection to maximize workload performance, which is critical database systems. In contrast existing methods, we seamlessly integrate recommendation rules and deep reinforcement learning, such that can recommend single-attribute multi-attribute indexes together for complex queries meanwhile support multiple-index access a table. Specifically, first propose five heuristic generate candidates. Then, formulate as learning task employ Deep Q Network (DQN) on it....

Although many updatable learned indexes have been proposed in recent years, whether they can outperform traditional approaches on disk remains unknown. In this study, we revisit and implement four state-of-the-art disk, compare them against the B+-tree under a wide range of settings. Through our evaluation, make some key observations: 1) Overall, performs well across workload types datasets. 2) A index could or other for specific workload. For example, PGM achieves best performance...

Customized 3D-printed structural parts are widely used in surgical robotics. To satisfy the mechanical properties and kinematic functions of these parts, a topology optimization technique is adopted to obtain optimal layout while meeting constraints objectives. However, currently faces some practical challenges that must be addressed, such as ensuring structures do not have significant defects when converted additive manufacturing models. address this problem, we designed 3D hierarchical...

Managing massive trajectory data from various moving objects has always been a demanding task. A desired system should be versatile in its supported query types and distance functions, of low storage cost, consistently efficient on processing different properties. Unfortunately, none the existing systems can meet above three criteria at same time. To this end, we propose VRE, v ersatile, r obust, e conomical system.VRE separates processing. In layer, novel segment-based model that takes...

Outdoor 3D object detection has played an essential role in the environment perception of autonomous driving. In complicated traffic situations, precise recognition provides indispensable information for prediction and planning dynamic system, improving self-driving safety reliability. However, with vehicle's veering, constant rotation surrounding scenario makes a challenge systems. Yet most existing methods have not focused on alleviating accuracy impairment brought by rotation, especially...

This paper introduces a kind of immune genetic algorithm which regards objective function as antigen, solution antibody and updates the population using evolutionary strategy. After economic emission load dispatch belongs to multi-objective constrained optimization problems is discussed, main processes solve this matter given. Through tests power system model with five coal-burning generating units, feasibility validity proved. And by comparing Hopfield neural network, quick constringency...

A hydraulic model is used to study the flowing process within a bottom-blowing furnace designed by company, and VOF adopted simulate its process. The method can describe formation, growing up separation actions in theorizing comparing simulation experimental results, this foundation of using research thermal furnace. It indicated that nozzle with disperse spouts stir larger zone, influence on four nozzles different structures. are benefit increase stirring reaction efficiencies, but pressure...

Query optimizer is at the heart of database systems. Cost-based studied in this paper adopted almost all current A cost-based introduces a plan enumeration algorithm to find (sub)plan, and then uses cost model obtain that plan, selects with lowest cost. In model, cardinality, number tuples through an operator, plays crucial role. Due inaccuracy cardinality estimation, errors huge space, cannot optimal execution for complex query reasonable time. paper, we first deeply study causes behind...

While in-memory learned indexes have shown promising performance as compared to B+-tree, most widely used databases in real applications still rely on disk-based operations. Based our experiments, we observe that directly applying the existing disk suffers from several drawbacks and cannot outperform a standard B+-tree cases. Therefore, this work make first attempt show how idea of index can benefit on-disk by proposing AULID, fully updatable achieve state-of-the-art across multiple workload...

The de-icing of a wing leading edge using an electro-impulse method benefits from its very low energy requirement and high efficiency. high-frequency mechanical vibration activated by electromagnetic pulse coil linked to impulse circuit fractures the ice accumulating on edge, is removed rapidly when flying. An improved criterion based transient dynamics employed accurately simulate (EIDI) process. To reduce computational expenses in modelling all rivet joints, simplified model structure...

With the development of self-attention mechanism, Transformer model has demonstrated its outstanding performance in computer vision domain. However, massive computation brought from full attention mechanism became a heavy burden for memory consumption. Sequentially, limitation consumption hinders deployment on embedded system where computing resources are limited. To remedy this problem, we propose novel economy named Couplformer, which decouples map into two sub-matrices and generates...

GBLUP, the most widely used genomic prediction (GP) method, consumes large and increasing amounts of computational resources as training population size increases due to inverse relationship matrix (GRM). Therefore, in this study, we developed a new method (RHEPCG) that avoids direct GRM by combining randomized Haseman-Elston (HE) regression (RHE-reg) preconditioned conjugate gradient (PCG). The simulation results demonstrate RHEPCG, cases, not only achieves similar predictive accuracy with...

Accurate potential energy surface (PES) calculation is the basis of molecular dynamics research. Using deep learning (DL) methods can improve speed PES while achieving competitive accuracy to <i>ab initio</i> methods. However, performance DL model extremely sensitive distribution training data. Without sufficient data, suffers from overfitting issues that lead catastrophic degradation on unseen samples. To solve this problem, based message passing paradigm graph neural networks, we...

In this paper, we study the problem of cardinality estimation for similarity search on high-dimensional data (CE4HD). We aim to perform CE4HD with high robustness (i.e., robust different datasets), query large variance and scale) efficiency. propose leverage selected objects (called reference objects) in database achieve above. Specifically, two techniques that adopt strategies select objects, as well support efficient computation dynamic databases. Extensive experiments datasets from...

Small file management is widely encountered in industrial areas. Consolidating small files can benefit the performance of data system. Many existing consolidation solutions fail to realize importance a proper schema. Therefore, they use very primitive and ineffective schemas. In this paper, we focus on proposing an effective robust Unlike most that only historical workload, consider workload uncertainty issue propose graph-clustering-based solution more future. To do this, introduce...

In order to investigate the mature law of Korla fragrant pear with mathematics method, pears in Aksu area be taken as study object, and fruit hardness soluble solid content (SSC) selected research index. The data two indices process are analyzed Excel, SigmaPlot Matlab. To draw change rule establish mathematical mode. shows that effective accumulated temperature reaching up 3000°C 3843°C is main stage pear, there an accelerating which. can precise description by established model provide...