In this paper, we study the optimal use of a conveying surface, called smart surface, using distributed microrobotic system that is designed to manipulate fragile micro objects. The smart surface is composed of a 2D array of decentralized micro-modules. Each micro-module is composed of a microactuator, a microsensor, computing and communication units. The cooperation among these micro-modules allows the micro-object to be accurately located and moved on the surface. We discuss in this paper the algorithmic solutions that allow the system to determine the best path to move an object from an initial to a target position in the surface. The optimality of the path is evaluated according to two different and complementary criteria corresponding to short and long term visions. The short term vision optimizes on the speed of conveyance. The long term vision optimizes on the lifespan of the system and maintenance issues. We describe the best way to deal with these two aspects and their impact on the smart surface performance. The results observed confirm the improvement of smart surface performance compared to a naive approach, with a gain in lifespan of up to 160% and good objects transfer times.