Sharing charging stations are an effective solution for daily usage of electric vehicles charging, however, the area with high demand cannot provide enough stations while there are plenty of stations left idle in remote areas with less demand. The core of the problem is the imbalance of demand and supply. In other word, we need to allocate the charging station to the appropriate locations to balance demand and supply. This study aims to solve the problem of locating charging stations for public electric vehicles (PUEVs), to improve the sharing charging level. We take into consideration the factors affecting charging station locations including mileage, PUEV distribution and passenger distribution. A Non-deterministic Polynomial (NP) model aiming to minimize the total vehicle service distance is developed. We use an agent-based model to simulate the optimized charging station location based on Anylogic. Through a case study of Beijing, we test the model in five situations. This paper concludes that priority, mileage, PUEV distribution and passenger distribution are the key factors affecting the location of PUEV charging stations, with exogenous variables such as the type of circuit and the voltage drawn as constants. The results of one situation show that the existing layout of the charging stations is unreasonable when charging frequency is sharply variant; this paper optimizes the existing location by improving the constraint for the smallest number of charging stations; the proposed model can be used for EV charging stations’ location in densely populated metropolis.
18 November 2015 in Information Technology and Management
High-speed train delay prediction has always been one of the important research issues in the railway dispatching. Accurate and interpretable delay prediction can enable staff to implement preventive measures and scheduling decisions in advance, and guide relevant departments to cooperate in completing complex transportation tasks, so as to improve rail transit operations, service quality, and the efficiency of train operation. This paper proposes a new interpretable model based on graph community neural network and time-series fuzzy decision tree. This model can well capture the influence of spatiotemporal characteristics, train community structure and multi-factor in high-speed train station delay prediction. Besides, the time series fuzzy decision tree based on multi-objective optimization and reduced error pruning can mine potential decision rules to improve the model's interpretability, transparency, and high reliability. Finally, we prove that the prediction effect of the proposed model is superior than the other 7 state-of-the-art models and our model is interpretable.
23 March 2019 in Advances in Production Engineering & Management
The gig economy has facilitated the growth of customized services through digital platforms that connect consumers with service providers. However, the surge in service providers has led to a “cold-start problem", which limits the effectiveness of personalized task recommendation systems. To address this challenge, this paper purposed a personalized recommendation system for human-centric consumer services in the gig economy. It addresses the problem by using meta-learning to generate suitable preference embeddings for workers with limited bidding history, interests, and working competence. The system includes a competence module with self-attention and interest modules to capture workers’ personalized preferences. The model is evaluated on real-world datasets from Freelancer.com, and the results demonstrate that it outperforms state-of-the-art models in accurately recommending suitable personalized tasks to both new and existing workers with skill-evolving. The proposed system can reduce task completion times and improve task quality by ensuring that tasks are assigned to the most suitable workers.