The reports presented by consulting firms show that annual energy costs generated by 340 million streetlights are expected to reach $23.9 to $42.5 billion by 2025. Those numbers reveal a motivation behind the research aiming at optimizing outdoor lighting energy efficiency. They show that even a small unit improvement can yield large benefits due to the effect of scale. The development of solid state lighting solutions enables highly effective modernization of street lighting installations. It allows obtaining power saving not only by replacing high pressure lamps with LEDs but also by improving a design quality and by introducing a dynamic street lighting control. Both methods, however, are not feasible for industry-standard software tools due to the significant complexity related to a configuration optimization, especially for large-scale projects. The goal of this article is presenting the workaround to the complexity issue, which is based on application of graph methods. They enable optimizing lighting installations in the scale of a city by providing a scalable computational environment. The presented case study shows that, thanks to applying the proposed method, one can design a lighting system which has the energy consumption reduced by up to 70%.