In engineering studies, finding the best route from a start point to an end point on pixel-based weighted maps is a big problem for researchers. For this problem many methods and algorithms have been developed until now. The "cost distance" (CD) and "cost path" (CP) tools that are used by a modified Dijkstra's algorithm and used by Environmental Systems Research Institute's (ESRI) ArcGIS Desktop 10 software are very fast and most preferred solutions for route optimization problems. Despite the advantages of these tools, they have the disadvantage of making a lot of curves with big angles. Especially in some engineering studies like power transmission lines' routing, the angle of the curves of the lines should not have big values. For overcoming this problem both genetic algorithm (GA) and artificial bee colony (ABC) algorithm that have been improved and adapted for the problem were used as powerful optimizers to find the best routes with fewer curves in those kinds of optimum route problems. New functions like smart direction sensing and improved random functions were developed for application of GA and ABC algorithms in power transmission lines routing studies. This study showed that the ABC algorithm's performance is better than GA. The accuracy of the algorithms was proven by comparing the results with the CD-CP tools' results. The experimental results showed that the improved algorithms gave better performance than Dijkstra's algorithm.