Robotic partition wall was realized as a space optimization system to assign variable volumes to two adjacent spaces with nonlinear and dynamic usage patterns. This prototype is built around 68HC11 microprocessor architecture from motorola combining sensor-systems for scanning human movements, user volumes, occupancy density etc, and motor mechanisms for kinetic response. The wall is equipped with low-level artificial neural network to model a week-long activity pattern and assign depending on varying volume requirements through out the week appropriate position for the partition wall. While in operation, the wall shifts laterally to make room for the upcoming function, shrinking the adjacent latent space. While conventional design practices are based around defining non-transformable fixed spacial volumes, robotic partition wall attempts to look for alternatives to this practice in the light of new digital technologies where spaces are defined dynamically.