Abstract Traffic prediction on road networks is highly challenging due to the complexity of traffic systems and a crucial task in successful intelligent system applications. Existing approaches mostly capture static spatial dependency relying prior knowledge graph structure. However, can be dynamic, sometimes physical structure may not reflect genuine relationship between roads. To better complex spatial‐temporal dependencies forecast conditions networks, multi‐step model named...

A critical problem in large neural networks is over parameterization with a number of weight parameters, which limits their use on edge devices due to prohibitive computational power and memory/storage requirements. To make more practical real-time industrial applications, they need be compressed advance. Since cannot train or access trained when internet resources are scarce, the preloading smaller essential. Various works literature have shown that redundant branches can pruned...

Solving the unit commitment (UC) problem in a computationally efficient manner has become increasingly crucial, especially context of high renewable energy penetration. This paper tackles this challenge by employing offline training model-free deep reinforcement learning (DRL) framework, thereby enhancing optimization efficiency UC problem. The complex modeling random variables is avoided reformulating as Markov decision process, where DRL-based method extracts knowledge regarding wind...

Peer-to-peer (P2P) energy trading offers a promising way for prosumers to achieve multi-bilateral trades, further aids the integration of distributed resources into distribution networks and facilitates low-carbon operation system. But realizing this potential requires overcoming challenges in model formulation optimization. This paper presents novel P2P framework carbon-aware based on carbon intensity analysis, where explicit emission cap constraints are embedded. To alleviate computational...

Self-similar growth and fractality are important properties found in many real-world networks, which could guide the modeling of network evolution anticipation new links. However, technology-convergence such characteristics have not yet received much attention. This study provides empirical evidence for self-similar field intelligent transportation systems. further investigates implications fractal link prediction via partial information decomposition. It is discovered that two different...

Blockchain, as an emerging technology and a disruptive innovation, has attracted attention from both academia industry. However, there are many potential risks associated with it, such the technical risk, legal risk privacy risk. A comprehensive analysis is crucial for cost-effective deployment of blockchain technology. Important adoption decisions, including when to deploy blockchain, how plan investment, transfer current businesses onto price service depend on this analysis. Yet very...

Data that house topological information is manifested as relationships between multiple variables via a graph formulation. Various methods have been developed for analyzing time series on the nodes of graphs but research works signals with volatility are limited. In this paper, we propose framework multivariate Generalized Autoregressive Conditional Heteroscedasticity (GARCH) models from spectral perspective Laplacian matrix. We introduce three graphical GARCH models: one symmetric Graph...

The over-parameterization of neural networks and the local optimality backpropagation algorithm have been two major problems associated with deep-learning. In order to reduce redundancy network parameters, conventional approach has prune branches small weights. However, this only solves problem parameter redundancy, not providing any global guarantees. paper, we overturn back-propagation combine sparse optimization weight using a non-convex method, namely Simulated Annealing. This method can...

Recently, the ever-increasing development of distributed energy resources has created new opportunities for active distribution networks. In particular, peer-to-peer (P2P) management systems facilitate coordination prosumers higher efficiency and flexibility. However, in P2P problems, Byzantine attacks may exist lead to nonconvergence infeasible solutions, computation performance needs be further improved. Therefore, a Byzantine-resilient approach is developed minimize overall cost...

Abstract Recently, generative models have been gradually emerging into the extended dataset field, showcasing their advantages. However, when it comes to generating tabular data, these often fail satisfy constraints of numerical columns, which cannot generate high-quality datasets that accurately represent real-world data and are suitable for intended downstream applications. Responding challenge, we propose a generation framework guided by task optimization (TDGGD). It incorporates three...

The main challenge in scheduling multi-cluster tools semiconductor manufacturing is the interactions among clusters. These create a k-unit optimal production that do not exist single-cluster tools. This paper analyzes of cycle with single-blade robots. A resource-based method used to analytically derive closed-form expressions for minimal time tool. Conditions decoupling and optimality conditions widely pull schedule are also presented. An example from industry illustrate derived formula conditions.

This study proposes a framework to diagnose stock market crashes and predict the subsequent price rebounds. Based on observation of anomalous changes in correlation networks during crashes, we extend log-periodic power-law model with metric that is proposed measure network anomalies. To calculate this metric, design prediction-guided anomaly detection algorithm based extreme value theory. Finally, hybrid indicator rebounds index by combining visibility graph-based model. Experiments are...

The hospital system is a complex service in which patients and hospitals interact make their decisions based on bounded rationality information. In this paper, we develop generative agent-based model to simulate the behavior of system. Our combines simulation queueing models mimic processes. goal understand growth size distribution hospitals. This includes agents for supply elements (i.e., with different resources expansion strategies) demand preferences selections) Three important questions...

This study analyzes the SARS-CoV-2 genome sequence mutations by modeling its nucleotide as a stochastic process in both time-series and spatial domain of gene sequence. In model, Markov Chain embedded Poisson random characterizes mutation rate matrix, while model delineates distribution inter-occurrence distances. Our experiment focuses on five key variants concern that had become global due to their high transmissibility virulence. The results reveal distinct asymmetries propensities among...

This paper proposes an educational game design framework named the Knowledge-based Personality, Emotion, and Action Cues (KPEAC). It aims to naturally enhance user's intrinsic motivation learning effectiveness by employing these novel Anthropomorphic Design (ADC). As example of this framework, we developed Chemical Life, a chemistry education game, where existing atoms molecules act as characters with assigned personality traits, emotions, reactions according chemical rules. For instance,...

Abstract Faced with dynamic and increasingly diversified public transport requirements, bus operators are urged to propose operational innovations sustain their competitiveness. In particular, ordinary operations heavily constrained by well-established route options, it is challenging accommodate passenger flows effectively a good level of resource utilization performance. Inspired the philosophy sharing economy, many available resources on road, such as minibuses private vehicles, can offer...

The imbalanced distribution of shared bikes in the dockless bike-sharing system (a typical example resource-sharing system), which may lead to potential customer churn and lost profit, gradually becomes a vital problem for firms their users. To resolve problem, we first formulate as Markovian queueing network with higher-demand nodes lower-demand nodes, can provide steady-state probabilities having certain number at one node. A model reduction method is then designed reduce complexity...

This paper considers the competition of service centers within a system. Each center is modeled as multiple-server queue. Service compete with each other in fair manner and adjust their resources (number servers) to accommodate demand obtained from competition. An agent-based approach used model adjustment processes. Our objective study size distribution under competition, i.e., answer question, will be uniform, normal, or skewed? It turns out that skewed (e.g., power law distribution)...