Search

Formulating optimal bidding strategies is pivotal for market participants to enhance electricity market profits. The main challenge for finding optimal bidding strategies is how to deal with system uncertainty, which stems from the inherent unpredictability and fluctuation within the electricity market. In the previous works, deep reinforcement learning (DRL) is proved a promising approach in multi-agent system with uncertainty. But few works model the relevance between agents for processing system uncertainty, especially the dynamic correlation in the operation of market. For this purpose, this paper proposes to model the correlation between agents to cope with the system uncertainty in a representative centralized double-sided auction market by combining graph convolutional neural network (GCN) with deep deterministic policy gradient (DDPG) algorithm, which is not only able to deal with the system uncertainty by aggregating correlative information of neighboring agents, but also helps obtain superior bidding strategies for the market participants. The proposed algorithm is evaluated on 4-bus, 30-bus and 57-bus congested network, where both supply side and demand side with elastic demand are modeled as RL agents. The results demonstrate that the proposed algorithm achieves higher system profits than the DRL based algorithms without GCN. • A bi-level strategic bidding model for market participants is established. • Reinforcement learning and graph convolutional neural network (GCN) are integrated. • GCN with adaptive adjacency matrix is learned to model the spatial correlation. • The temporal correlation of historical nodal price is modeled. [ABSTRACT FROM AUTHOR]

• The present study proposed a novel graph convolution model to forecast future traffic speeds. • The proposed model differentiated the intensity of connecting to neighbor roads unlike existing GCNs. • The present study was focused on devising a GCN model that mimic true propagation patterns of traffic. • The proposed model shows promise for application to a real-time traffic forecasting system. The traffic state in an urban transportation network is determined via spatio-temporal traffic propagation. In early traffic forecasting studies, time-series models were adopted to accommodate autocorrelations between traffic states. The incorporation of spatial correlations into the forecasting of traffic states, however, involved a computational burden. Deep learning technologies were recently introduced to traffic forecasting in order to accommodate the spatio-temporal dependencies among traffic states. In the present study, we devised a novel graph-based neural network that expanded the existing graph convolutional neural network (GCN). The proposed model allowed us to differentiate the intensity of connecting to neighbor roads, unlike existing GCNs that give equal weight to each neighbor road. A plausible model architecture that mimicked real traffic propagation was established based on the graph convolution. The domain knowledge was efficiently incorporated into a neural network architecture. The present study also employed a generative adversarial framework to ensure that a forecasted traffic state could be as realistic as possible considering the joint probabilistic density of real traffic states. The forecasting performance of the proposed model surpassed that of the original GCN model, and the estimated adjacency matrices revealed the hidden nature of real traffic propagation. [ABSTRACT FROM AUTHOR]