Decentralized Reinforcement Learning Approach for Microgrid Energy Management in Stochastic Environment

Table 2

The average results of EMS based on decentralized reinforcement learning over 800 days.

Scenario	Daily generated/consumed energy (kWh)		Daily profit/expense (rial)
	I	II	I	II
Wind	125.2	124.2	40561	51681
PV	70.2	70.1	31226	41279
Diesel	63	117.4	14670	28274
Fuel cell	75.2	141.7	17153	41539
Microturbine	75.3	142.1	24401	52317
Battery	10.7	1.2	−2312	91
Electrical load	618.9	216.6	81184	28049
Thermal load	257.1	99	60282	22849
Main grid	475.7	−211	436720	−46247