Task Offloading with Power Control for Mobile Edge Computing Using Reinforcement Learning-Based Markov Decision Process

Algorithm 1

Q-learning algorithm with -greedy policy.

	Initialization:
	; , ; , and ;
	while ( is the is the maximum number of iterations)
	for each UE in
	if exploration
	chooses an action a arbitrarily with probability , ;
	else exploitation
	chooses an action ;
	end if
	perform a and get a reward and a successor state ;
	update the Q-value function according to equation (15);
	;
	end for
	t  t + 1;
	end while