Imitation Learning Based on Deep Reinforcement Learning for Solving Scheduling Problems

Abstract

The paradigm shift in urban planning toward citizen participation originates from the Smart City concept, as politicians and scientists argue that citizens should be included in the design of their environment. This led to the development of urban participation platforms and was enhanced by the COVID-19 pandemic as on-site participation was unavailable. Past projects showed that urban participation platforms can reach thousands of citizens, but it became apparent that citizens' contributions vary widely and are sometimes not understandable and comprehensible which limits their value for urban projects. Therefore, we examined how an AI-based feedback system can increase citizens' argumentation on urban platforms. For this, an explorative comparison of two prototypes was conducted by applying Argumentation Theory and Mayring's qualitative content analysis to empirically analyze collected data. The findings highlight that the developed AI-based feedback system supports citizens and leads to more argumentative and comprehensible argumentations on urban participation platforms.