Personalized Route Recommendation With Neural Network Enhanced Search Algorithm

Abstract

In this work, we study an important task in location-based services, namely Personalized Route Recommendation (PRR). Given a road network, the PRR task aims to generate user-specific route suggestions for replying to users' route queries. A classic approach is to adapt search algorithms to construct pathfinding-like solutions. These methods typically focus on reducing search space with suitable heuristic strategies. For these search algorithms, heuristic strategies are often handcrafted, which are not flexible to work in complicated task settings. In addition, it is difficult to utilize useful context information in the search procedure. To develop a more principled solution to the PRR task, we propose to improve search algorithms with neural networks for solving the PRR task based on the widely used A * algorithm. The main idea of our solution is to automatically learn the cost functions in A * algorithms, which is the key of heuristic search algorithms. Our model consists of two main components. First, we employ attention-based Recurrent Neural Networks (RNN) to model the cost from the source to the candidate location by incorporating useful context information. Instead of learning a single cost value, the RNN component is able to learn a time-varying vectorized representation for the moving state of a user. Second, we propose to use an estimation network for predicting the cost from a candidate location to the destination. For capturing structural characteristics, the estimation network is built on top of position-aware graph attention networks. The two components are integrated in a principled way for deriving a more accurate cost of a candidate location for the A * algorithm. Extensive experiment results on three real-world datasets have shown the effectiveness and robustness of the proposed model.

Robust Road Network Representation Learning

“…Trajectories, representing the movement of vehicles within a city, are a crucial data source to provide supplementary insights for tasks related to road networks [27]. In particular, road networks explicitly impose structural constraints that govern the traversal of trajectories, forming the foundations for applications such as route planning [1], [2], anomaly detection [28] and destination prediction [29]. Conversely, trajectories provide rich knowledge of traveling semantics for road networks [17], which can effectively enhance tasks that may not necessarily involve trajectory data.…”

“…R OAD networks, as a fundamental yet indispensable component in transportation systems, play a crucial role in various downstream transport planning tasks. These tasks include trajectory-based tasks like route inference [1], [2] and road segment-based tasks like traffic forecasting [3], [4]. Recent studies have increasingly focused on deriving effective representations that can capture the intrinsic characteristics of road networks.…”

A dynamic bidirectional heuristic trust path search algorithm

“…In order to solve the efficiency problem in path search, many researchers [12][13][14][15] introduce heuristic information to search, which can make search focus on the target and prune some ineffective or impossible paths to reduce workload. Among the path search algorithms, A* algorithm has been widely used in global path planning [16][17][18]. It can also be used to the area [19][20][21][22] of recommendation system, medicine, image processing etc.…”

Self-supervised Representation Learning for Trip Recommendation

“…Trip recommendation is significantly different from traditional route planning in that the latter is usually to search for the shortest or a minimum cost-based route. In contrast, trip recommendation takes more into consideration the diversity and personalized needs of a POI sequence, which is also not constrained by the road network (Wang, Wu, and Zhao 2021). Therefore, trip recommendation is a more challenging task in location-based recommendation services.…”

“…• DeepTrip(Gao et al 2021): It is an end-to-end trip recommendation method, while using a GAN-style neural network to approximate the similarity between the query and the trip. • NASR+(Wang, Wu, and Zhao 2021): It is a deep neural network-based A* algorithm for route planning constrained by the road network. Since the check-in data is not constrained by the road network, we only use its RNN module enhanced by the attention mechanism for a fair comparison.…”