Effective caching of paths and regions for mobile services

Abstract

Cheap GPS-enabled mobile devices have made online location based services (LBSs) increasingly common. Examples of LBS are: route planner, navigation assistance, restaurant recommendation, meeting point recommendation, and nearby friend notification. In location based services, caching can be used to reduce computational load or communication latency for queries. In this thesis, we consider three problems on caching paths and regions for mobile services. Our first problem is to investigate caching techniques for shortest paths. However, existing caching techniques do not exploit the optimal substructure property of shortest paths. We propose a novel caching problem for shortest paths, and develop algorithms and data structures for this problem. Our second problem is to explore trade-offs between the lengths and the query coverage of concise shortest paths, in online driving direction services. Driving instructions are equivalent to so-called concise shortest paths, which occupy much less space than the corresponding shortest paths. The caching of concise shortest paths has two opposite effects on the cache hit ratio. The cache can accommodate a larger number of concise paths, but each individual concise path contains fewer nodes and so may answer fewer shortest path queries. We formulate the concept of a generic shortest path enabling trade-off between a paths size and the number of queries it can answer. We present algorithms for computing and caching generic concise shortest paths. Our third problem is to compute safe regions for the sum-optimal meeting point notification problem. It has applications like social networking services or online games, in which multiple moving users in a group may wish to be continuously notified about the best meeting point from their locations. To reduce the communication frequency, an application server can compute safe regions, which capture the validity of query results with respect to users locations. Unfortunately, the safe regions in our problem exhibit characteristics such as irregular shapes and inter-dependencies, which render existing methods for single safe region inapplicable. We present algorithms for computing safe regions for sum-optimal meeting point. In summary, we make the following contributions: (i) algorithms and data structures for caching of shortest paths; (ii) algorithms for computing generic concise shortest paths; (iii) algorithms for computing the safe regions for the sum-optimal point notification problem.