在Python 3.8，Pygame 1.9.6 下运行会报错：
Traceback (most recent call last):
File "D:/RL/MaCA/fight_mp.py", line 14, in
from interface import Environment
File "D:\RL\MaCA\environment\interface.py", line 15, in
from world.em_battle import BattleField
File "D:\RL\MaCA\environment\world_init_.py", line 1, in
pyarmor(name, file, b'\xec\x50\x8c\x64\x26\x42\xd6\x01\x10\x54\xca\x9c\xb6\x33\x81\x05\x30\xdb\x46\xb3\x6d\x0c\xa2\x2d\x19\xe2\x8f\xf9\x44\x25\xd5\x53\xc7\x67\x06\x76\x71\x8d\x42\xe2\xc3\xef\x00\x16\xa3\x91\x11\x67\x09\x92\x9a\x72\x18\xf6\x31\xbd\x95\x78\x69\xe0\x84\xc1\x30\x40\x93\x62\xc4\x8e\xd6\xf0\x0d\x5b\x82\x29\x55\x63\xd9\x3e\x7a\x4e\x40\xcb\xae\x0e\x9d\x45\x72\xa4\x22\x50\xc5\x00\x06\x2d\x29\xe5\xe1\x5f\xa6\xee\x04\xfd\x9f\xc1\x10\xe5\xd5\x07\xef\xb6\xee\x36\xcc\x39\x5c\xa0\xab\x6f\x6e\x9e\xb6\xf5\x42\x3f\x41\xeb\x1b\x0a\x07\xbe\x67\x3e\x0f\x8f\xe7\x61\x8a\xcf\x9b\xa3\x22\x07\xb0\x4c\xc0\x53\x6f\xf9\x0c\x1c\x61\xfa\xa0\x28\xa4', 1)
MemoryError
涉及加密代码，没法自行更改。
注：在Python 3.6，Pygame 1.9.6 测试没有问题。

我在环境中尝试了fighter10v10与fighter5v5，使用了simple dqn算法与自己编写的ippo算法与fix rule agent进行对抗，实际动作空间仅选取了航向和攻击，并且我降低了航向的个数，但效果都很差。
能否提供一些RL训练的trick？雷达开关频点动作是否对结果有很大的影响呢（是否是因为雷达频点固定为1而导致无法探测到敌方？）？是否需要添加更多非稀疏奖励来引导agent学习？

希望能得到回答，谢谢

您好，目前在学习这个程序，能跑，只是可视化渲染一直出不来，是哪个代码块需要修改吗？mac，感谢

你好，我将“environment/”文件夹设置为“Sources Root”，也把fight_mp.py的“Working Directory”设置为MaCA根目录了，但是运行的时候报错。请问这是什么原因？

Traceback (most recent call last):
File "E:/MaCA-master/fight.py", line 15, in
from interface import Environment
File "E:\MaCA-master\environment\interface.py", line 4, in
pyarmor_runtime()
File "E:\MaCA-master\environment\pytransform.py", line 274, in pyarmor_runtime
pyarmor_init(path)
File "E:\MaCA-master\environment\pytransform.py", line 271, in pyarmor_init
init_pytransform()
File "E:\MaCA-master\environment\pytransform.py", line 72, in wrap
raise PytransformError(format_message(errmsg, *args, **kwargs))
pytransform.PytransformError: b"\nOpen pyshield license file 'E:\\xe9\xa1\xb9\xe7\x9b\xae\\xe6\x99\xba\xe8\x84\x91\MaCA-master\environment\pyshield.lic'\nNo such file or directory\nInitialize license key failed\nInitialize pytransform failed"

Process finished with exit code 1

你好，我在运行程序时回出现这样的错误，，请问您这边知道如何处理吗

XXX lineno: 41, opcode: 160
Traceback (most recent call last):
File "/home/simsimi/MaCA/fight.py", line 77, in
random_pos=args.random_pos, log=log_flag)
File "/home/simsimi/MaCA/environment/interface.py", line 44, in init
self.map = Map(map_path)
File "", line 41, in init
SystemError: unknown opcode

Hi there
I was able to train a model on GPU
And the model was fine when inference on GPU

But my GPU server does not have display access so I wanted to show the demo on my local CPU, but when I ran the code I met this problem

And as you know the files under environment are all encrypted so I have no idea what when wrong, can you shed some lights on me ? Thanks A lot !!!!!

您好，我自学RL有一段时间了，因为对海战感兴趣，一直想把RL应用在海战上，但是gym上没有合适的环境，一直想自己写一个，但是又无从下手，看了您的这个项目虽有所启发，但您的项目有些复杂，我还是有太多不理解的地方，不知您是否有时间指导一下呢？我前期想写一个一条舰艇防导弹攻击的场景，算法方面看的资料也不少了，就是环境这块不知该怎么搞。我现在在研究您的environment，如果您有时间指导学生一下那万分感谢了。

Traceback (most recent call last):
File "E:/code/github/MaCA/fight_mp.py", line 14, in
from interface import Environment
File "E:\code\github\MaCA\environment\interface.py", line 15, in
from world.em_battle import BattleField
File "E:\code\github\MaCA\environment\world\em_battle.py", line 1, in
pyarmor(name, file, b'...', 1)
File "", line 27, in
File "E:\code\github\MaCA\environment\world\position_calc.py", line 1, in
pyarmor(name, file, '...', 1)
File "", line 26, in
ModuleNotFoundError: No module named 'numpy'

Traceback (most recent call last):
File "Y:/demo/MaCA/fight_mp.py", line 14, in
from interface import Environment
File "Y:\demo\MaCA\environment\interface.py", line 4, in
pyarmor_runtime()
File "Y:\demo\MaCA\environment\pytransform.py", line 274, in pyarmor_runtime
pyarmor_init(path)
File "Y:\demo\MaCA\environment\pytransform.py", line 271, in pyarmor_init
init_pytransform()
File "Y:\demo\MaCA\environment\pytransform.py", line 72, in wrap
raise PytransformError(format_message(errmsg, *args, **kwargs))
pytransform.PytransformError: b'License is expired\nCheck license failed\nInitialize pytransform failed'

Hi, I am using python 3.6.3 (in pyenv and venv) on Ubuntu 18.04 LTS in terminal. The fight_mp works fine, but when I run fight.py --map=1000_1000_fighter10v10 --agent1 fix_rule --agent2 simple --log --log_path default_log I get the 'Error: agent simple model data not exist!' error. I tried to run main.py from train folder first, but I get 'ModuleNotFoundError: No module named 'agent'' error. Generally when I try to run files, it seems like it can't load objects from other files. I tried adding files'
path to the python path, but it didn't work.
I would like to ask you if you have any solution to this problem and also if you could clarify the order that i should run the files (for training and running the DQN case.) Thank you very much in advance.

D:\Conda\envs\fly\python.exe D:/pycharm_workplace/MaCA-master/train/simple/main.py
pygame 2.0.1 (SDL 2.0.14, Python 3.6.2)
Hello from the pygame community. https://www.pygame.org/contribute.html
Traceback (most recent call last):
File "D:/pycharm_workplace/MaCA-master/train/simple/main.py", line 45, in
env = Environment(MAP_PATH, red_agent_obs_ind, blue_agent_obs_ind, render=RENDER)
File "D:\pycharm_workplace\MaCA-master\environment\interface.py", line 53, in init
side2_name=side2_name)
File "", line 128, in init
File "", line 29, in init
FileNotFoundError: No such file or directory.

Process finished with exit code 1
在运行main.py的时候出现了这个问题，感觉是加密的模块读取失误？想请您帮忙看看
另外，想请问license什么时候过期，我想在毕设的时候引用一下贵司的平台，担心到时候license过期了…

很多地方被加密了，有些无法调用

很多地方都加密了，想看看一下源码学习学习

用系统自带的main和dqn函数训练时，训练到一定阶段当我方有战机损失时，程序报错，list out of range，是怎么回事

已经按照要求配置了需要的环境，但是在运行示例的main.py时报内存错误，试了8G和16G内存的电脑，都不行，是pytorch版本的问题？

RuntimeError: $ Torch: not enough memory: you tried to allocate 3GB. Buy new RAM! at ..\aten\src\TH\THGeneral.cpp:201

根目录更改成了environment 并且main的working directory也改成了maca的根目录 但是一直出现报ModuleNotFoundError: No module named 'agent'
请问是怎么回事 有解决方案吗

如题

大家好，我在安装完之后运行python fight_mp.py中的“from interface import environment”时出现如下错误，请问是什么原因造成的。

Hi, dose the environment support algorithm as A3C(Asynchronous Advantage Actor-Critic)?
or support parallel training net ?

And how do I cite this environment in my paper? May I list the authors names in reference?

不知道是什么原因，windows 10 下不能完全看到整张地图，因此我想在一个更小的地图上进行训练。但是新建了新的地图图片以及.map文件，并在main里修改了map-path，提示找不到.map文件，我想是不是因为无法对environment里的loadmap进行修改造成的错误。
所以我想知道是否有可能重建地图，或者在win10上看到地图全部？
第二就是除了核心模型外，是不是科技将地图以及其他可扩展的功能开放?让我们可以根据需要自己修改，我认为这是一个比较好的开源平台，如果只是局限在比赛的内容上就有些局限了。还是期待您可以开源一个类似的完全开放的平台。

文档描述每一个fighter的striking_list是一个元素为字典的列表，实际代码中是元素为int型的列表。
striking_list
打击列表[{'attacker_id': 攻击者编号, 'target_id': 目标编号, 'missile_type':导弹类型(1：远程， 2：中程), information_source_type: 信息来源类型（0： L 探测单元， 1:S 探测单元， 2：战机， 3：被动侦测） step_count:step 计数, op_count: 观测点计数, del_ind: 删除指示(默认 false)}]

实际输出是：[3,3]

运行fight_mp.py后提示
File "fight_mp.py", line 14, in
from interface import Environment
ModuleNotFoundError: No module named 'interface'

您好，可以告知观测数据字典中的 'info' 和 'screen'分别指的是什么吗？