本文會介紹打磚塊基本(Non-Machine-Learning)，本題材取自於 MLGame.py，https://github.com/LanKuDot/MLGame?fbclid=IwAR3Eni0s0wp-Wpv-2iLB8Vtt_Bn8l3TBw6lDhXFh3WYhl5EQ342Yk_yTYpc 此參考中，先以利用不是產生 Machine-Learning，來完成打磚塊的程式，相信大家清楚打磚塊規則，以磚塊全數消失就是過關為主，如果球低於板子則是失敗，所以我們第一次先不使用，Machine-Learning的做法來完成自動使用板子來打球體的動作。

import games.arkanoid.communication as comm
from games.arkanoid.communication import ( \
    SceneInfo, GameInstruction, GameStatus, PlatformAction
)

def ml_loop():

    ball_position_history = []

    comm.ml_ready()
    while True:
        
        scene_info = comm.get_scene_info()
        ball_position_history.append(scene_info.ball)
        platform_center_x = scene_info.platform[0]+20
        if (len(ball_position_history)) == 1:
            ball_going_down = 0
        elif ball_position_history[-1][1] - ball_position_history[-2][1]:
            ball_going_down = 1
            vy = ball_position_history[-1][1]-ball_position_history[-2][1]
            vx = ball_position_history[-1][0]-ball_position_history[-2][0]
        else:
            ball_going_down =0

        if scene_info.status == GameStatus.GAME_OVER or \
            scene_info.status == GameStatus.GAME_PASS:

            comm.ml_ready()
            continue

        if ball_going_down ==1 and ball_position_history[-1][1] >= 150 :

            ball_destination = ball_position_history[-1][0]+(390-ball_position_history[-1][1]/vy)*vx
            #print(platform_center_x)

            print(ball_destination)
            if ball_destination >= 195 :
                ball_destination = 195-(ball_destination - 195)
                #print(ball_destination)

                comm.send_instruction(scene_info.frame, PlatformAction.MOVE_RIGHT)

            elif ball_destination <= 0 :
                ball_destination = -ball_destination
                #print(ball_destination)
                comm.send_instruction(scene_info.frame, PlatformAction.MOVE_LEFT)

        else:
            ball_destination = platform_center_x

上方程式碼，是利用 cmd 指令 python MLGame.py -r arkanoid 2 -i ml_play_template.py 進行驅動上方指令動作，vy = ball_position_history[-1][1]-ball_position_history[-2][1]
vx = ball_position_history[-1][0]-ball_position_history[-2][0]
以上兩條式子，是將球的參數算出速度以及距離，根據這個距離資料來做判斷來藉由讓板子移動。

以上為不專業的AI介紹，謝謝各位的觀看，那我們下篇見~~~~