Journal of Surveying and Mapping Engineering Sept. 2014, Vol. 2 Iss. 3, PP. 56-64 - 56 - On the Automation of the Cartographic Insetting Procedure with Parallel Numerical Searching Algorithm Papakonstantinou Apostolos *1 , Christos Christodoulou 2 , Varsamis Dimitris 3 1 Garto GI Lab, Geography Department, Aegean University, Lesvos, Greece 2,3 Department of Informatics, Technological Educational Institute of Central Macedonia, Serres, Greece *1 apapak@geo.aegean.gr; 2 xristodoulou223@gmail.com; 3 dvarsam@teiser.gr Abstract-Insetting procedure is a time consuming problem in island cartography, where cartographers dealing with special cartographic problems confronted in the portrayal of island regions, and demands the use of specially developed algorithms and software tools. This paper presents the implementation of a parallel numerical searching algorithm for the optimization of the insetting cartographic procedure and showcases the use of this parallel algorithmic procedure to tackle the computational challenge of such problems. Experiments with mathematical representation datasets and real-world datasets showed that the parallel algorithm drastically reduced the execution time and achieved fairly high speed-ups and efficiencies. Experiments also showed the algorithm scaled reasonably well as the number of processors increased and as the problem size increased. Keywords- Inset Map; Geovisualization; Parallel Computing; Numerical Algorithm I. INTRODUCTION Cartography is the study and practice of making maps. It involves all the appropriate steps to transform the geographical information into suitable geovisualizations (maps) in static or dynamic form. Cartography depends on sound geographical knowledge of the surface being mapped, as well as the many skills and instrumental tools to the mapping process [1]. The cartography often encounters problems that require specially developed software tools. One of these problems faced by cartographers is the insetting procedure. This problem occurs strongly when mapping islands. One of the most commonly faced problems is the need of inset map creation for very small islands, and sometimes isolated ones, that must be displayed in the main map. The key factor for insetting in Island Cartography is the “complexity of land discontinuity” comprised of a number of islands that vary in size and shape creating a complex area surrounded by sea [2]. It is very common for small and big size islands to coexist, thus rendering representation at the same scale is problematic. The scale of the main map often causes small-sized islands not to be clearly visible by the map-reader or not to be clearly drawn [3, 4]. The special developed software was the Inset Mapper (IM). IM is a software tool for automated place and scale selection for inset maps that overcomes these procedure difficulties and is developed in MATLAB, along with other pertinent cartographic functionalities. This software tool attempts to emulate the steps a cartographer takes when deciding where to place and in which scale within a map to inset. To meet these requirements, IM employs some new strategies of cartographic representation for the inset placement procedure, and for scale selection, in multiple inset maps. Varsamis et al. presented in [5] a parallel processing approach, implemented in a quad core PC, of the searching algorithm for the insetting procedure in order to reduce the calculation time of the computational procedure and the total execution time of the IM software tool, respectively. This paper presents the implementation of a parallel numerical searching algorithm using a grid computer machine with 64 cores for the optimization of the insetting cartographic procedure and showcases the use of this parallel algorithmic procedure to tackle the computational challenge of such problems. A computational procedure following specific cartographic rules for calculating the appropriate position and the optimal scale for an inset was presented [6, 7]. Parallel processing is one of the most successful efforts to introduce explicit parallelism to high level programming languages [8-10]. This approach is selected because many useful computations can be conducted in terms of a set of independent sub-computations, each strongly associated to an element of a large dataset, where such computations are inherently parallelizable. Parallel processing is particularly convenient because it can scale easily to large problem sizes [9, 10], such as searching into large geographic datasets. The aim of this paper is to present a different approach for the problem of the insetting procedure in order to reduce the calculation time and to find the optimal solution in insetting cartographic procedure. In particular in section 2, is presented the parallelization and the implementation of the numerical searching algorithm. Additionally, in this section, is explained how the numerical algorithm works. Finally in section 3, the implemented algorithm is applied using the parallel computing toolbox of Matlab [11-13] and a comparison of execution times for specific geographic datasets is conducted.