Artificial intelligent traffic signal controlling system, method and computer program product thereof

A traffic signal controlling system includes a pre-training neural network, a decision neural network and a control unit. The pre-training neural network executes reinforcement learning by using a plurality of training data. The decision neural network analyzes a real time traffic data according to training result of the pre-training neural network to generate a decision for controlling a traffic signal within a controllable range. The control unit controls the traffic signal within the controllable range according to the decision.