| Summary: | 碩士 === 逢甲大學 === 資訊工程學系 === 105 === In this study, we formulated a novel route planning query referred to as the Patrolling Skyline Path Query. The term patrolling road refers to situations in which vehicles patrol a particular area and follow particular roads/routes according to a set schedule. For example, roads that require snowplow service are referred to as Patrolling roads. However, identifying these roads can be difficult, due to changes in the accumulation of snow over time and the need to take into account distance and time costs as well as issues related to safety. No existing search mechanism is able to deal with these kinds of variables. The following hard points must be addressed. (1) How can we identify patrolling roads efficiently? (2) How can we boost query speeds? (3) How can we minimize the number of detours?
In this study, we developed two methods to address these issues. The first method uses an index model to reduce the time required to identify roads that have not been serviced for a extended period. We also developed a sub-network of roads to reduce the number of nodes and edges and thereby enhance the efficiency of conventional skyline path queries while prioritizing roads of importance to the largest number of users and taking into account the tolerance of users for detours. We also developed a variation of the proposed method based on the concept of spatial skyline queries to enable the exclusion of some roads from the patrol schedule in order to reduce the overall workload. Simulation results demonstrate that the second method is more efficient with regard to time costs and the accuracy of results.
|
|---|
