Multi Criteria Decision Making for n-Dimensional Vertical Partitions

Abstract

The Skyline processing is being used for multi criteria decision making. They are being widely used for retrieving better data points from a given set of data points according to the client preferences. Here we assume that the data is being vertically partitioned among several servers and also there is skyline in all possible subspaces. We take up data in 3-Dimensions x, y and z. It is based on the assumption that no two points have the same co-ordinates on any of the axis x, y or z. Here we first find the point “Pdom” which is the dominating point that eliminates a large number of records and hence the data is pruned in each of the server. Later we find the local region of dominance among the servers in which the data has been vertically decomposed. From the entire local dominance region the global region of dominance is found. The data points that come under the global region are not transmitted to the client. The union-intersection operation is performed over the various pruned regions among the servers and hence the final sets of points that do not lie within the pruned region are transmitted to the client. Here we have proposed an algorithm called SNDVP (Skyline for n-Dimensional Vertical Partitions) that uses the r function which is used to obtain the local region of dominance in n-dimensions.