Tropical cyclones (TCs), some of the most influential weather events across the globe, are modulated by the El Niño–Southern Oscillation (ENSO). However, little is known about the feedback of TCs on ENSO. Here, observational and modelling evidence shows that TC activity in the southeastern western North Pacific can affect the Niño-3.4 index 3 months later. Increased TC activity in July–September can significantly contribute to the intensity of ENSO in October–December by weakening the Walker circulation and enhancing eastward-propagating oceanic Kelvin waves in the tropical Pacific. Thus, the greater the accumulated cyclone energy, the stronger (weaker) the El Niño (La Niña). A new physics-based empirical model for ENSO is constructed that significantly outperforms current models in predicting ENSO intensity from July to December and addressing the problem about the target period slippage of ENSO. Results suggest that TCs may provide significant cross-scale feedback to ENSO.

This paper investigates the impact of the South China Sea summer monsoon (SCSSM) on the Indian Ocean dipole (IOD). The results show that the SCSSM has a significant positive relationship with the IOD over the boreal summer [June–August (JJA)] and fall [September–November (SON)]. When the SCSSM is strong, the enhanced southwesterly winds that bring more water vapor to the western North Pacific (WNP) lead to surplus precipitation in the WNP, inducing anomalous ascending there. Consequently, the anomalous descending branch of the SCSSM Hadley circulation (SCSSMHC) develops over the Maritime Continent (MC), favoring deficit precipitation in situ. The precipitation dipole over the WNP and MC as well as the enhanced SCSSMHC leads to intensification of the southeasterly anomalies off Sumatra and Java, which then contributes to the negative sea surface temperature (SST) anomalies through the positive wind–evaporation–SST and wind–thermocline–SST (Bjerknes) feedbacks. Consequently, a positive IOD develops because of the increased zonal gradient of the tropical Indian Ocean SST anomalies and vice versa. The SCSSM has a peak correlation with the IOD when the former leads the latter by three months. This implies that a positive IOD can persist from JJA to SON and reach its mature phase within the frame of the positive Bjerknes feedback in SON. In addition, the local and remote SST anomalies in the tropical Indian and Pacific Oceans have a slight influence on the relationship between the SCSSM and precipitation dipole over the WNP and MC.

The rapid development of wireless communications brings a tremendous increase in the amount number of data streams and poses significant challenges to the traditional routing protocols. In this paper, we leverage deep reinforcement learning (DRL) for router selection in the network with heavy traffic, aiming at reducing the network congestion and the length of the data transmission path. We first illustrate the challenges of the existing routing protocols when the amount of the data explodes. We then utilize the Markov decision process (RSMDP) to formulate the routing problem. Two novel deep Q network (DQN)based algorithms are designed to reduce the network congestion probability with a short transmission path: one focusing on reducing the congestion probability; while the other focuses on shortening the transmission path. The simulation results demonstrate that the proposed algorithms can achieve higher network throughput comparing to existing routing algorithms in heavy network traffic scenarios. INDEX TERMSDeep reinforcement learning, routing, network congestion, network throughput, deep Q network.

No abstract

No abstract

No abstract

No abstract

Abstract. An efficient simulation execution engine is crucial for agentbased traffic simulation. Depending on the size of the simulation scenario the execution engine would have to update several thousand agents during a single time step. This update may also include route calculations which are computationally expensive. The ability to dynamically re-calculate the route of agents is a feature often not required in classical microscopic traffic simulations. However, for the agent-based traffic simulation which is part of the Scalable Electro-Mobility Simulation (SEMSim) platform, the routing ability of agents is an important feature. In this paper, we describe a multi-threaded simulation engine that explicitly supports routing capabilities for every agent. In addition, we analyse the efficiency and performance of our execution model in the context of a Singapore-based simulation scenario.