Three cost models are derived to measure Data Broadcast Wait (DBW), Data Access Time in the multiplexing scheme (ATDataMul) where both data and indices are broadcast in the same channel, and Data Access Time in the separate channel scheme (ATDataSep) where data and indices are broadcast in two separate channels. Hypergraph representations are used to represent the spatial relationships of both point data and graph data. The broadcast data placement problem is then converted to the graph layout problem. A framework for classifying ordering heuristics for different types of geographical data is presented. A low-polynomial cost approximation graph layout method is used to solve the DBW minimization problem. Based on the proven monotonic relationship between ATData Sep and DBW, the same approximation method is also used for AT DataSep optimization. A novel method is developed to optimize ATDataMul. Experiments using both synthetic and real data are conducted to evaluate the performance of the ordering heuristics and optimization methods. The results show that R-Tree traversal ordering heuristic in conjunction with the optimization methods is effective for sequencing point data for spatial range query processing, while graph partition tree traversal ordering heuristic in conjunction with the optimization methods is suitable for sequencing graph data for network path query processing over air.

Geographical data broadcasting is suitable for many large scale dissemination-based applications due to its independence of number of users, and thus it can serve as an important part of intelligent information infrastructures for modern cities. In broadcast systems, query response time is greatly affected by the order in which data items are being broadcast. However, existing broadcast ordering techniques are not suitable for geographical data because of the multi-dimension and rich semantics of geographical data. This research develops cost models and methods for placing geographical data items in a broadcast channel based on their spatial semantics to reduce response time and energy consumption for processing spatial queries on point data and graph data.