In this section, MORSE-E is our energy-efficient selfoptimizing scheduler, which is naturally derived using our proposed automatic reward derivation and multi-factor feature selection procedures, using Et 2 as the objective function and the same 50 candidate state attributes used in Section 6.