Efficient skyline query processing in wireless sensor networks

Enhanced Distributed Dynamic Skyline Query for Wireless Sensor Networks

“…Su et al [30] proposed an algorithm, known as Skyline Sensor Algorithm (SkySensor), in a customized DCS method in order to collect and store all sensor readings and retrieve skyline results efficiently from the network. The major disadvantages of SkySensor are: (1) SkySensor needs increased effort from an application designer who has to write the center location of each cluster in such a way that no two clusters overlap; (2) number of clusters in a sensor network depends on the number of attributes of a tuple, therefore, an active sensor and actuator network, with higher rate of data generation and a lower number of attributes leads to high concentration of data in a small portion of the network which creates congestion and a hot spot around the cluster and limits the ultimate goal or advantage of DCS; and (3) in the case of resilience to node failure, an inefficient and old technique referred to as local replication is used, which is expected to incur high storage cost and increased data loss during a node group failure.…”

Histogram Estimation for Optimal Filter Skyline Query Processing in Wireless Sensor Networks

“…There are two filter methods, one is tuple filter approach and another is grid filter approach. Su et al [10] proposed a skyline sensor algorithm (SkySensor) to efficiently retrieve skyline results from a sensor network. A cluster-based architecture is designed in SkySensor to collect all sensor readings from sensor nodes.…”

“…(1) ← sensor numbers (2) for = 1 to do (3) Initialization: LS( ) = 0 ⊳ initializing each node's local skyline; (4) sort tuples in node( ) in ascending order of the first dimension (5) for each tuple in node( ) do (6) for each tuple in LS( ) do 7if is not dominated by then (8) insert to LS( ) ⊳ insert the tuple to the skyline (9) end if (10) end for (11) end for (12) (LS( )) ⊳ compute the dominance probability of each local skyline tuple (13) ← tuples with the biggest dominance probability (14) if ( ) > 1 then (15) DR( ) ⊳ compute the dominance region of every tuple in (16) ( ) ← tuple with the largest value of DR (17) end if (18) end for (19) ( ( )) ⊳ use the in-network approach to find tuple with best dominance ability (20) ( ) ⊳ broadcast the filter to the network (21) for each tuple in LS( ) do (22) if is dominated by then (23) remove from LS( ) (24) end if (25) end for (26) (LS( )) ⊳ use the in-network approach to compute the final skyline Algorithm 1: HFA algorithm. Finally, the tuple which makes (9) and (10) the largest is chosen as final filter tuple flt .…”

A novel lightweight iteration filtering algorithm research in wireless sensor networks

“…In [15] the author proposes an adaptive algorithm, which can continuously minimize the communication cost, it adopts Client/Server framework, maintaining the skyline of dynamic objects. In [16] the author proposes a skyline algorithm based on the data-centric storage strategy, and two prune strategies were proposed to prune the nodes and tuples that have no possibility of being the final results. In [17] the author proposes an energy efficient optimized routing structure based algorithm, considering optimizing the data communication from the routing layer.…”