A segmented parallel expansion algorithm for keyword-aware optimal route query

Keyword-aware Optimal Route Query (KOR) searches for an optimal route with the shortest traveling time under the conditions of full coverage of keywords and route budget, and is a high-frequency query in numerous map applications. Shortening the execution time is the significant goal of KOR optimization. The state-of-the-art algorithms primarily utilize various route expansion approaches to evaluate KORs, and focus on pruning strategies to reduce the search scale and shorten the execution time. Those strategies are effective in controlling the search scale for short routes, however, ineffective for long routes, because the search scale increases exponentially with the search depth. Therefore, this paper proposes PSE-KOR, a segmented parallel expansion algorithm for KOR, to address the issue for long routes. PSE-KOR constructs the routes with keyword vertexes as necessary passing nodes to satisfy the full coverage of keywords and budget, and divides the route into multiple segments taking the keyword vertexes as the boundary to limit the search scale and expands them in parallel to accelerate execution. For each route segment, a local budget limit pruning strategy is proposed to constrain the expansion direction and search depth, while reducing the interference among multiple segments. Extensive experiments verify the efficiency and effectiveness of PSE-KOR.