A decision-making approach using point-cloud-based granular information

Real-life decision problems are often solved by using a very limited set of data points. The computational complexity of such decision problems can be handled easily by using a mathematical entity called stochastic point cloud (SPC). In general, SPC is a form of granular information and is simulated by a data-driven stochastic process. In this study, particular interests are given on normal-distribution-driven SPCs. When the parameters of a normally distributed variable is not clearly known due to the lack of information or due to the multiplicity of regression analysis, it forms a set of SPCs. Three types of such SPCs are described in detail in this study. The effectiveness of such SPCs in solving a real-life decision problem (i.e., how to minimize vehicle emissions in Abu Dhabi Emirate of United Arab Emirates) is also shown. To develop more realistic and man-machine-friendly decision support systems one can use SPCs. (C) 2010 Elsevier B.V. All rights reserved.