A novel technique to design a fuzzy logic controller using Q(λ)-learning and genetic algorithms in the pursuit-evasion game

Q(λ)-learning adaptive fuzzy logic controllers for pursuit-evasion differential games

“…The state and action spaces are discrete and their corresponding value function is stored in what is known as a Q-table. To use Q-learning with continuous systems (continuous state and action spaces), one can discretize the state and action spaces [21,[31][32][33][34][35][36] or use some type of function approximation such as FISs [26][27][28], NNs [12,19,37], or use some type of optimization technique such as genetic algorithms [38,39].…”

“…We test the proposed technique in three cases. In case 1, the evader is constrained to use a simple control strategy (as the one used in [35]). The control strategy that is used by the evader is to run away from the pursuer along the LOS.…”

Learning and design of fuzzy logic controllers for pursuit-evasion differential games

“…In this chapter, we propose a novel technique based on Q(A)-learning and GAs to tune the input and the output parameters of FLC automatically. The proposed technique is called Q(\) -learning based genetic fuzzy logic controller (QLBGFLC) [67,68]. The proposed technique is applied to three versions of pursuit-evasion differential games.…”

Learning in pursuit-evasion differential games using reinforcement fuzzy learning

“…Instead, one can use a "function approximator" such as fuzzy systems to represent the continuous state space and action space [5,8,10,11].…”

“…Observe the received immediate reward r at the subsequent state s'. The simplest form of Q-learning called one-step Q-learning is given by Q(s,a) = Q(s, a) + a(r + 7 maxQ(s', a') -Q{s, a)) (4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14) where r is the received immediate reward at time t +1, a; is the learning rate and 7 is the discount factor. The part r + 7 max" Q{s', a') -Q(s, a) in (4.14) is the temporaldifference error calculated at time t + 1.…”