In order to evaluate our approach, we chose the scenario of a warehouse where an agent equipped with our intelligent reasoning system (further called intelligent agent) has to fetch randomly placed items while avoiding further agents. For representing such scenarios, we use a grid world consisting of discrete cells like illustrated in Figures
\ref{213840},
\ref{841210}. Our intelligent agent can move around in its world via single actions for moving to the north, east, south, or west from its current cell. That is, if it is not blocked by a wall or another agent occupying the target cell (then it would remain in the current cell). Technically, we implemented our scenario as an extension to the OpenAI gym environment gym-maze (see
https://github.com/MattChanTK/gym-maze).