Network space intelligent game decision-making method and system

The invention discloses a network space intelligent game decision-making method and system. The method comprises the following steps: setting a game space of a network space intelligent game decision-making model and initializing the game space; establishing a mapping relationship between the agent reinforcement learning model and the network space intelligent game decision model, and interacting the network space agent with the game space based on the mapping relationship; in the game of one round, the attack agent and the defense agent act in each game round according to own strategies, and the attack agent and the defense agent play the game for multiple times until the number of game steps reaches a set threshold value; and the attack agent and the defense agent perform a game of multiple rounds, when the game reaches the maximum round number and the maximum game step number of each round, the network space intelligent game is terminated, and a winning party is determined according to a game result and a