This study aims to explore how cross-border e-commerce (CBEC) enterprises achieve a desired balance state of supply chain resilience and vulnerability. proposes the concept vulnerability, which refers proximity current ideal one anti-ideal one. To evaluate state, model is constructed using integrating fuzzy Analytic Hierarchy Process (AHP) Technique for Order Preference by Similarity Ideal Solution (TOPSIS) methods, holistically consider drivers vulnerability (CBECSC). An empirical included...

A deep learning (DL) model is inherently imprecise. To address this problem, existing techniques retrain a DL over larger training dataset or with the help of fault injected models using insight failing test cases in model. In paper, we present Apricot, novel weight-adaptation approach to fixing iteratively. Our key observation that if architecture trained many different subsets original dataset, weights resultant reduced (rDLM) can provide insights on adjustment direction and magnitude...

This paper studies a repeated game between manufacturer and two competing suppliers with imperfect monitoring. We present principal-agent model for managing long-term supplier relationships using unique form of measurement incentive scheme. measure supplier's overall performance rating equivalent to its continuation utility (the expected total discounted future payoffs), incentivize effort larger allocations business. obtain the vector suppliers' ratings as state Markov decision process, we...

To reduce the impact of wildfires on operation power systems, a back-propagation neural network (BPNN) model is used to evaluate wildfire risk distribution after feature selection. Data from 14 types wildfire-related features, including anthropogenic, geographical, and meteorological factors, were collected public data websites local departments. The weight ranking was calculated using filtering wrapper methods form five subsets. These are as input sets BPNN training, parameters optimized by...

The paper aims to optimize the functional area layout of Weihai cold chain logistics base. On basis positioning base, optimizes and determines relative position each interval by design specific functions At same time, system planning method (SLP) is used analyze correlation degree relationship nonlogistics interval, explore possibility optimal By constructing model international using SLP design, clarifying points principles area, putting forward corresponding scheme, so as make facility...

We present the new predictor-corrector methods for systems of nonlinear differential equations, based on method exponential time differencing. compare schemes with explicit multistep differencing and Adams–Bashforth–Moulton method. The numerical results show that are more accurate efficient than Adams has been developed perfected by studies.

Task selection (picking an appropriate labeling task) and worker (assigning the task to a suitable worker) are two major challenges in assignment for crowdsourcing. Recently, has been successfully addressed by bandit-based (BBTA) method, while not thoroughly investigated yet. In this paper, we experimentally compare several strategies borrowed from active learning literature, show that least confidence strategy significantly improves performance of

Green supply chain is one of the important means to achieve sustainable development automobile industry. This paper aims improve current evaluation index system green management in industry, so as evaluate environmental impact, resource utilization efficiency and sustainability whole process The improved can provide a basis for long-term enterprises level their own chain, also an tool government degree In future, with progress protection requirements needs be further improved.

It is widely accepted that land use and cover (LULC) an important conditioning factor for landslide occurrence, especially when considering the role of tree roots in stabilizing slopes consolidating soil. However, it still difficult to assess impacts a specific LULC type on distribution. The objective present study reveal relationship between bamboo distribution at regional scale. We aim answer following question: do areas covered by have higher susceptibility landslides? Wenzhou City SE...

We consider a non-parametric penalized likelihood approach for model building called basis pursuit (LBP) that determines the probabilities of binary outcomes given explanatory vectors while automatically selecting important features. The LBP involves parameters balance competing goals maximizing log-likelihood and minimizing terms. These are selected to minimize proxy misclassification error, namely, randomized, generalized approximate cross validation (ranGACV) function. ranGACV function is...

Zhang, H.; Tang. L., 2019. A Nash equilibrium game model for seafood safety regulation. In: Gong, D.; Zhu, H., and Liu, R. (eds.), Selected Topics in Coastal Research: Engineering, Industry, Economy, Sustainable Development. Journal of Research, Special Issue No. 94, pp. 818–822. Coconut Creek (Florida), ISSN 0749-0208.To prevent the incidents from becoming serious, it is vital to find root causes accidents. This paper expatiates trading patterns online-sold resulting situations. Then a...

This research establishes a measurement model to solve the problems exist in collaboration of supply chain alliances, especially between suppliers and manufacturers. paper firstly identifies 21 indicators including profit level management manufacturers through expert brainstorming method determines weight each indicator by experts scoring. Then authors get index (C), comprehensive evaluation (T) obtain degree (D) Primary, Transition, Intermediate, Good Excellent. At last, do empirical...

Developing a deep learning (DL) based software system is slow. One of the critical issues to conduct many trials and errors in developing DL model that usually serves as major component such system. A reason for this inefficiency progress gradual reduction gap between under training ground truths. Prior techniques commonly focus on optimizing after have formed. They are insensitive how dataset provided batches, making their approaches nonproactive deal with errors. In article, we propose...

The accuracy of DL models may not meet the user's expectations. To tackle this problem, existing work proposed diverse approaches, such as using more optimized training processes and samples to evolve model structure or parameters faulty models. In paper, we present Plum, a novel hyperheuristic approach fixing deep learning Plum generates set candidates by applying low-level repair strategies. It then evaluates prioritizes strategies based on their overall effects exhibited these outputs...

Executing program paths outside the ones tested means that is executing scenarios not before deployment. No existing technique can produce a precisely contains an arbitrary set of in each procedure program. This paper presents first work, novel called OPE, to address this problem. OPE builds transformed target for every It extends with additional branches and basic blocks code include all remaining given procedure. The resultant functionally equivalent achieves inherent strict separation...

For the tumor gene expression profile data that aiming to high-dimension small samples, how select classification feature of samples among thousands genes effectively is difficult problems for analysis on profile.First partition set into K average divisions, use Lasso method performing selection each respectively, and then merge selected division subset together perform feather again, get final gene.This experiment adopts Support Vector Machine (SVM) as classifier, take performance by Leave...