On the complexity of parallel coordinate descent

In this work we study the parallel coordinate descent method (PCDM) proposed by Richtárik and Taká? [Parallel coordinate descent methods for big data optimization, Math. Program. Ser. A (2015), pp. 1–52] for minimizing a regularized convex function. We adopt elements from the work of Lu and Xiao [On the complexity analysis of randomized block-coordinate descent methods, Math. Program. Ser. A 152(1–2) (2015), pp. 615–642], and combine them with several new insights, to obtain sharper iteration complexity results for PCDM than those presented in [Richtárik and Taká?, Parallel coordinate descent methods for big data optimization, Math. Program. Ser. A (2015), pp. 1–52]. Moreover, we show that PCDM is monotonic in expectation, which was not confirmed in [Richtárik and Taká?, Parallel coordinate descent methods for big data optimization, Math. Program. Ser. A (2015), pp. 1–52], and we also derive the first high probability iteration complexity result where the initial levelset is unbounded.     

Orthorhombic‐pseudocubic phase transition and piezoelectric properties of (Na0.5K0.5)(Nb1−xSbx)‐SrZrO3 ceramics

0.96(Na_0.5K_0.5)(Nb_1?xSb_x)‐0.04SrZrO₃ ceramics with 0.0≤x≤0.06 were well sintered at 1060°C for 6 hours without a secondary phase. Orthorhombic‐tetragonal transition temperature (T_O‐T) and Curie temperature (T_C) decreased with the addition of Sb₂O₅. The decrease in T_C was considerable compared to that in T_O‐T, and thus the tetragonal phase zone disappeared when x exceeded 0.03. Therefore, a broad peak for orthorhombic‐pseudocubic transition as opposed to that for orthorhombic‐tetragonal transition appeared at 115°C‐78.2°C for specimens with 0.04≤x≤0.06. An orthorhombic structure was observed for specimens with x≤0.03. However, the polymorphic phase boundary structure containing orthorhombic and pseudocubic structures was formed for the specimens 0.04≤x≤0.06. Furthermore, a specimen with x=0.055 exhibited a large piezoelectric strain constant of 325 pC/N, indicating that the coexistence of orthorhombic and pseudocubic structures could improve the piezoelectric properties of (Na_0.5K_0.5)NbO₃‐based lead‐free piezoelectric ceramics.     

Suppressing Lithium Dendrite Growth by Metallic Coating on a Separator

Lithium (Li) metal is one of the most promising candidates for the anode in high‐energy‐density batteries. However, Li dendrite growth induces a significant safety concerns in these batteries. Here, a multifunctional separator through coating a thin electronic conductive film on one side of the conventional polymer separator facing the Li anode is proposed for the purpose of Li dendrite suppression and cycling stability improvement. The ultrathin Cu film on one side of the polyethylene support serves as an additional conducting agent to facilitate electrochemical stripping/deposition of Li metal with less accumulation of electrically isolated or “dead” Li. Furthermore, its electrically conductive nature guides the backside plating of Li metal and modulates the Li deposition morphology via dendrite merging. In addition, metallic Cu film coating can also improve thermal stability of the separator and enhance the safety of the batteries. Due to its unique beneficial features, this separator enables stable cycling of Li metal anode with enhanced Coulombic efficiency during extended cycles in Li metal batteries and increases the lifetime of Li metal anode by preventing short‐circuit failures even under extensive Li metal deposition.     

Development of a Data‐Driven Framework for Real‐Time Travel Time Prediction

Travel time prediction is one of the most important components in Intelligent Transportation Systems implementation. Various related techniques have been developed, but the efforts for improving the applicability of long‐term prediction in a real‐time manner have been lacking. Existing methods do not fully utilize the advantages of the state‐of‐the‐art cloud system and large amount of data due to computation issues. We propose a new prediction framework for real‐time travel time services in the cloud system. A distinctive feature is that the prediction is done with the entire data of a road section to stably and accurately produce the long‐term (at least 6‐hour prediction horizon) predicted value. Another distinctive feature is that the framework uses a hierarchical pattern matching called Multilevel k‐nearest neighbor (Mk‐NN) method which is compared with the conventional k‐NN method and Nearest Historical average method. The results show that the method can more accurately and robustly predict the long‐term travel time with shorter computation time.     

Survival of Shigella sonnei on smooth tomato surfaces, in potato salad and in raw ground beef

This study investigated the survival of a five-strain Shigella sonnei cocktail on smooth tomato surfaces, in potato salad and in raw ground beef. All inocula were resistant to the antibiotic rifampicin to allow simple detection of the target culture among the indigenous microflora of the food samples. Inoculated tomatoes were stored at 13 degrees C/85% relative humidity, the standard holding conditions for mature, green tomatoes. Inoculated potato salad and ground beef samples were stored at 2.5 degrees C and 8.0 degrees C to study the effects of varied refrigerated temperatures. Surviving populations were estimated using a three-tube most probable number (MPN) method. Tryptic soy broth tubes supplemented with 100 ppm rifampicin were inoculated with appropriate dilutions of the recovered inocula and scored for growth after overnight enrichment. S. sonnei populations declined rapidly to undetectable levels (2 days) when dried on smooth surfaces of tomatoes. S. sonnei populations did not decrease in potato salad and ground beef stored at 2.5 degrees C and 8.0 degrees C over the shelf-life of the products.     

A stochastic load balancing algorithm for i‐Computing

This paper presents a stochastic dynamic load balancing algorithm for Internet computing, which is a new type of distributed computing involving heterogeneous workstations from different organizations on the Internet. To realize the practical environment, we assume the system to be comprised of heterogeneous, untrusted and non‐dedicated workstations connected by a non‐dedicated network. Our algorithm uses the product of the average processing time and the queue length of system jobs as the load index. Dynamic communication delay is included in the execution cost calculation. The transfer policy and the location policy are combined in a stochastic algorithm. State information exchange is done via information feedback and mutual updating. Simulations demonstrate that our algorithm outperforms conventional approaches over a wide range of system parameters. These results are reconfirmed by empirical experiments after we have implemented the algorithms on the Distributed Java Machine global virtual machine. Copyright © 2003 John Wiley & Sons, Ltd.