fix: small changes for temperature

fix: correct calculation of max possible reward per step in inference script

fix: update return values in reset to be like step result as openenv expects it

feat: add inference script for smart grid environment