Circle-Covering Algorithm for Multiple-POI Searching

Abstract

Nowadays, there is a tendency that the Location-based service (LBS) is becoming more and more popular. That is because LBS can provide customized services according to user's current location information. What is more, it can easily recommend the information of a point of interest (POI) to the user. When a user queries a series of POIs, LBS can return a cluster of these POIs. This paper aims to find a given-sized circle that can cover the maximum total weights of POIs for which people require. The weight of a POI reflects its quality, hence, our aim is to help people find the best POIs nearby. In this paper, every cluster is considered as a point. We stipulate an order to traverse these points and update the maximum weight. The challenge of our proposed algorithm lies in the "cast effect", which means the most of the returned POI clusters may be discarded when some hot POIs are blocked. To overcome this challenge, our proposed algorithm will reduce the impact of the cast effect when calculating the weight of a circle. The simulation and the conclusion are presented at the end of this paper.