While I was running some simulation, this happened. I cannot replicate it since it happened only one. However, you might want to investigate it

terminate called after throwing an instance of 'std::runtime_error'
  what():  /home/MAgent/src/gridworld/Map.cc:60 : cannot find a blank position in a filled map

XXXX-virtual-machine:~/MAgent-master$ sudo apt-get install cmake libwebsocketpp-dev
正在读取软件包列表... 完成
正在分析软件包的依赖关系树
正在读取状态信息... 完成
E: 未发现软件包 libwebsocketpp-dev
前两个软件包没有问题 ，但是libwebsocketpp-dev显示未发现，应该如何解决。

Hey, thanks for publishing this great platform!

Question: During --train mode, sampling seems to stop, if there's no agent of e.g. g_f left. Any tips how to prevent this / implement many different types of agents which get consumed one by one?

Use case: I tried to add food with different types of value to train_gather.py to see if agents gather the more valuable food first. Unfortunately I'm not able to set this up, because sampling simply stops when one of the food agents has been used up.

    g_s = cfg.add_group(agent)
    g_f = cfg.add_group(food)
    g_f_2 = cfg.add_group(food)

    a = gw.AgentSymbol(g_s, index='any')
    b = gw.AgentSymbol(g_f, index='any')
    c = gw.AgentSymbol(g_f_2, index='any')

    cfg.add_reward_rule(gw.Event(a, 'attack', b), receiver=a, value=5.0)
    cfg.add_reward_rule(gw.Event(a, 'attack', c), receiver=a, value=1.0)

Thank you!

您好，我按readme操作，在Ubuntu14.04上运行bash build.sh ，总是会报错。报错内容如下

/usr/lib/gcc/x86_64-linux-gnu/5/include/smmintrin.h: In function ‘__m128i _mm_cvtepi16_epi64(__m128i)’:
/usr/lib/gcc/x86_64-linux-gnu/5/include/smmintrin.h:508:59: error: cannot convert ‘__v8hi {aka short int}’ to ‘__vector(8) short int’ for argument ‘1’ to ‘__vector(2) long long int __builtin_ia32_pmovsxwq128(__vector(8) short int)’
   return (__m128i) __builtin_ia32_pmovsxwq128 ((__v8hi)__X);
                                                           ^
/usr/lib/gcc/x86_64-linux-gnu/5/include/smmintrin.h: In function ‘__m128i _mm_cvtepi8_epi16(__m128i)’:
/usr/lib/gcc/x86_64-linux-gnu/5/include/smmintrin.h:514:60: error: cannot convert ‘__v16qi {aka char}’ to ‘__vector(16) cha^Cmake[2]: *** wait: 没有子进程。 停止。

目前我的cmake版本是3.2.2，g++版本是5.4.1。我有在谷歌检索类似内容，但都没有答案。初步推断是g++的版本问题。因而想向您咨询g++版本。

Hi, at first, I want to express my thanks to this awesome framework!

I want to obtain the neighbors of each agent, i.e., the spacial local view include the observed agents of both group 1 and group 2, I want to obtain the IDs and positions of all it's neighbors.

I'm not sure if ubuntu 14 is supported for this project, because when I sudo apt-get install libwebsocketpp-dev, there was no matching package.

When I build:

The C compiler identification is GNU 4.8.4
-- The CXX compiler identification is GNU 4.8.4
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Configuring done
-- Generating done
-- Build files have been written to: /home/shawnlue/Applications/MAgent/build
Scanning dependencies of target magent
Scanning dependencies of target render
Scanning dependencies of target testlib
[  3%] [  6%] [ 10%] Building CXX object CMakeFiles/render.dir/src/render/backend/websocket.cc.o
Building CXX object CMakeFiles/magent.dir/src/runtime_api.cc.o
[ 13%] Building CXX object CMakeFiles/magent.dir/src/temp_c_booster.cc.o
Building CXX object CMakeFiles/testlib.dir/src/runtime_api.cc.o
In file included from /home/shawnlue/Applications/MAgent/src/render/backend/websocket.cc:1:0:
/home/shawnlue/Applications/MAgent/src/render/backend/websocket.h:4:53: fatal error: websocketpp/config/asio_no_tls_client.hpp: 没有那个文件或目录
 #include <websocketpp/config/asio_no_tls_client.hpp>
                                                     ^
compilation terminated.
make[2]: *** [CMakeFiles/render.dir/src/render/backend/websocket.cc.o] 错误 1
make[1]: *** [CMakeFiles/render.dir/all] 错误 2
make[1]: *** 正在等待未完成的任务....
[ 16%] Building CXX object CMakeFiles/testlib.dir/src/temp_c_booster.cc.o
[ 20%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/RenderGenerator.cc.o
[ 23%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/GridWorld.cc.o
/home/shawnlue/Applications/MAgent/src/gridworld/GridWorld.cc:456:0: warning: ignoring #pragma omp declare [-Wunknown-pragmas]
     #pragma omp declare reduction (merge : std::vector<RenderAttackEvent> : omp_out.insert(omp_out.end(), omp_in.begin(), omp_in.end()))
 ^
[ 26%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/Map.cc.o
[ 30%] Building CXX object CMakeFiles/magent.dir/src/gridworld/RenderGenerator.cc.o
/home/shawnlue/Applications/MAgent/src/gridworld/GridWorld.cc: In member function ‘virtual void magent::gridworld::GridWorld::step(int*)’:
/home/shawnlue/Applications/MAgent/src/gridworld/GridWorld.cc:474:40: error: expected ‘+’, ‘*’, ‘-’, ‘&’, ‘^’, ‘|’, ‘&&’, ‘||’, ‘min’ or ‘max’ before ‘merge’
     #pragma omp parallel for reduction(merge: render_attack_buffer)
                                        ^
make[2]: *** [CMakeFiles/testlib.dir/src/gridworld/GridWorld.cc.o] 错误 1
make[2]: *** 正在等待未完成的任务....
[ 33%] Building CXX object CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o
/home/shawnlue/Applications/MAgent/src/gridworld/GridWorld.cc:456:0: warning: ignoring #pragma omp declare [-Wunknown-pragmas]
     #pragma omp declare reduction (merge : std::vector<RenderAttackEvent> : omp_out.insert(omp_out.end(), omp_in.begin(), omp_in.end()))
 ^
/home/shawnlue/Applications/MAgent/src/gridworld/GridWorld.cc: In member function ‘virtual void magent::gridworld::GridWorld::step(int*)’:
/home/shawnlue/Applications/MAgent/src/gridworld/GridWorld.cc:474:40: error: expected ‘+’, ‘*’, ‘-’, ‘&’, ‘^’, ‘|’, ‘&&’, ‘||’, ‘min’ or ‘max’ before ‘merge’
     #pragma omp parallel for reduction(merge: render_attack_buffer)
                                        ^
[ 36%] Building CXX object CMakeFiles/magent.dir/src/gridworld/Map.cc.o
[ 40%] Building CXX object CMakeFiles/magent.dir/src/gridworld/AgentType.cc.o
make[2]: *** [CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o] 错误 1
make[2]: *** 正在等待未完成的任务....
make[1]: *** [CMakeFiles/testlib.dir/all] 错误 2
make[1]: *** [CMakeFiles/magent.dir/all] 错误 2
make: *** [all] 错误 2

Hope someone can help me to fix this.

for python3，you need change the urllib2 in utility.py to urllib.request :)

I met this error in the moring ,but i did not copy it at that time.
Here are some pictures about it.One is occured when running in the terminal , another is occored when runnning in pycharm.

When I run the examples the out-of-memory errors occured,And how to tune infer_batch_size smaller in models?

Hi geek-ai:

Thank you for contributing this great work!
I want know how MAgent handle collision?
Could you provide some information about the mechanism of avoiding collision or handling the collision when it happens?

Thank you for your time!

Hello,I want to know the framework is weather wrote by C++?And what are the files .cc in the file of src?

I have read the code and have some questions about the action label. In the doc file (get_start.md) you have provided, there are 23 different actions agents can take. However, I have no idea about how to map the action label to a specific action (left or right,1 step or 2 step,etc). I would be appreciate if you can provide more details.

How watch to the result of example?For the pursuit ,gathering and battle.And that result where is it saved?

Hello,

We were checking the implementation of DQN in multi-agent environment (pursuit) as given in this repo. Everything seemed the same as the usual DQN, except for the input views (observations) that are being passed for each agent. What exactly do they represent and why does it have 4 dimensions ? As per my understanding in the usual DQN we only have observations with 3 dimensions.

Also, please let me know if there are any other differences compared to the single-agent DQN.

Thanks!

briandeMacBook-Pro:MAgent brian$ brew install --HEAD david-icracked/websocketpp/websocketpp
==> Installing websocketpp from david-icracked/websocketpp
==> Cloning git://github.com/zaphoyd/websocketpp
Updating /Users/brian/Library/Caches/Homebrew/websocketpp--git
==> Checking out branch master
Error: Calling Formula#std_cmake_parameters is disabled!
Use Formula#std_cmake_args instead.
/usr/local/Homebrew/Library/Taps/david-icracked/homebrew-websocketpp/websocketpp.rb:9:in `install'
Please report this to the david-icracked/websocketpp tap!
Or, even better, submit a PR to fix it!

system:
Ubuntu 16.04.3 LTS
conda 4.4.7
Python 2.7.14 :: Anaconda, Inc.
gcc (GCC) 4.8.5
libjsoncpp-dev is already the newest version (1.7.2-1).
libboost-system-dev is already the newest version (1.58.0.1ubuntu1).
libwebsocketpp-dev is already the newest version (0.7.0-5).
cmake is already the newest version (3.5.1-1ubuntu3).

(tensorflow_p27) ubuntu@ip-172-31-24-200:~/cs/MAgent$ bash build.sh 
-- The C compiler identification is GNU 4.8.5
-- The CXX compiler identification is GNU 4.8.5
-- Check for working C compiler: /home/ubuntu/anaconda3/envs/tensorflow_p27/bin/cc
-- Check for working C compiler: /home/ubuntu/anaconda3/envs/tensorflow_p27/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /home/ubuntu/anaconda3/envs/tensorflow_p27/bin/c++
-- Check for working CXX compiler: /home/ubuntu/anaconda3/envs/tensorflow_p27/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Configuring done
-- Generating done
-- Build files have been written to: /home/ubuntu/cs/MAgent/build
Scanning dependencies of target magent
[  3%] Building CXX object CMakeFiles/magent.dir/src/temp_c_booster.cc.o
[  6%] Building CXX object CMakeFiles/magent.dir/src/runtime_api.cc.o
[  9%] Building CXX object CMakeFiles/magent.dir/src/gridworld/RenderGenerator.cc.o
[ 12%] Building CXX object CMakeFiles/magent.dir/src/gridworld/AgentType.cc.o
[ 15%] Building CXX object CMakeFiles/magent.dir/src/gridworld/RewardEngine.cc.o
[ 18%] Building CXX object CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o
/home/ubuntu/cs/MAgent/src/gridworld/GridWorld.cc:456:0: warning: ignoring #pragma omp declare [-Wunknown-pragmas]
     #pragma omp declare reduction (merge : std::vector<RenderAttackEvent> : omp_out.insert(omp_out.end(), omp_in.begin(), omp_in.end()))
 ^
/home/ubuntu/cs/MAgent/src/gridworld/GridWorld.cc: In member function ‘virtual void magent::gridworld::GridWorld::step(int*)’:
/home/ubuntu/cs/MAgent/src/gridworld/GridWorld.cc:474:40: error: expected ‘+’, ‘*’, ‘-’, ‘&’, ‘^’, ‘|’, ‘&&’, ‘||’, ‘min’ or ‘max’ before ‘merge’
     #pragma omp parallel for reduction(merge: render_attack_buffer)
                                        ^
CMakeFiles/magent.dir/build.make:182: recipe for target 'CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o' failed
make[2]: *** [CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o] Error 1
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/magent.dir/all' failed
make[1]: *** [CMakeFiles/magent.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2
Scanning dependencies of target render
Scanning dependencies of target testlib
[  3%] Building CXX object CMakeFiles/render.dir/src/render/backend/websocket.cc.o
[  6%] Building CXX object CMakeFiles/render.dir/src/render/backend/render.cc.o
[  9%] Building CXX object CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o
[ 12%] Building CXX object CMakeFiles/testlib.dir/src/temp_c_booster.cc.o
/home/ubuntu/cs/MAgent/src/gridworld/GridWorld.cc:456:0: warning: ignoring #pragma omp declare [-Wunknown-pragmas]
     #pragma omp declare reduction (merge : std::vector<RenderAttackEvent> : omp_out.insert(omp_out.end(), omp_in.begin(), omp_in.end()))
 ^
/home/ubuntu/cs/MAgent/src/gridworld/GridWorld.cc: In member function ‘virtual void magent::gridworld::GridWorld::step(int*)’:
/home/ubuntu/cs/MAgent/src/gridworld/GridWorld.cc:474:40: error: expected ‘+’, ‘*’, ‘-’, ‘&’, ‘^’, ‘|’, ‘&&’, ‘||’, ‘min’ or ‘max’ before ‘merge’
     #pragma omp parallel for reduction(merge: render_attack_buffer)
                                        ^
CMakeFiles/magent.dir/build.make:182: recipe for target 'CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o' failed
make[2]: *** [CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o] Error 1
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/magent.dir/all' failed
make[1]: *** [CMakeFiles/magent.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
[ 15%] Building CXX object CMakeFiles/render.dir/src/render/backend/data.cc.o
[ 18%] Building CXX object CMakeFiles/testlib.dir/src/runtime_api.cc.o
[ 21%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/RenderGenerator.cc.o
[ 24%] Building CXX object CMakeFiles/render.dir/src/render/backend/text.cc.o
[ 27%] Building CXX object CMakeFiles/render.dir/src/render/backend/utility/logger.cc.o
/home/ubuntu/cs/MAgent/src/render/backend/utility/logger.cc: In member function ‘void magent::render::Logger::log(const string&)’:
/home/ubuntu/cs/MAgent/src/render/backend/utility/logger.cc:14:20: error: ‘put_time’ is not a member of ‘std’
         ostream << std::put_time(std::localtime(&time), "[%Y-%m-%d %H:%M:%S] [")
                    ^
CMakeFiles/render.dir/build.make:158: recipe for target 'CMakeFiles/render.dir/src/render/backend/utility/logger.cc.o' failed
make[2]: *** [CMakeFiles/render.dir/src/render/backend/utility/logger.cc.o] Error 1
make[2]: *** Waiting for unfinished jobs....
[ 30%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/AgentType.cc.o
[ 33%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/RewardEngine.cc.o
[ 36%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/GridWorld.cc.o
/home/ubuntu/cs/MAgent/src/gridworld/GridWorld.cc:456:0: warning: ignoring #pragma omp declare [-Wunknown-pragmas]
     #pragma omp declare reduction (merge : std::vector<RenderAttackEvent> : omp_out.insert(omp_out.end(), omp_in.begin(), omp_in.end()))
 ^
[ 39%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/test.cc.o
[ 42%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/Map.cc.o
/home/ubuntu/cs/MAgent/src/gridworld/GridWorld.cc: In member function ‘virtual void magent::gridworld::GridWorld::step(int*)’:
/home/ubuntu/cs/MAgent/src/gridworld/GridWorld.cc:474:40: error: expected ‘+’, ‘*’, ‘-’, ‘&’, ‘^’, ‘|’, ‘&&’, ‘||’, ‘min’ or ‘max’ before ‘merge’
     #pragma omp parallel for reduction(merge: render_attack_buffer)
                                        ^
CMakeFiles/testlib.dir/build.make:182: recipe for target 'CMakeFiles/testlib.dir/src/gridworld/GridWorld.cc.o' failed
make[2]: *** [CMakeFiles/testlib.dir/src/gridworld/GridWorld.cc.o] Error 1
make[2]: *** Waiting for unfinished jobs....
CMakeFiles/Makefile2:141: recipe for target 'CMakeFiles/testlib.dir/all' failed
make[1]: *** [CMakeFiles/testlib.dir/all] Error 2
CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/render.dir/all' failed
make[1]: *** [CMakeFiles/render.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2

I am getting this message. How do I fix this?
I am using AWS Deep Learning AMI. So I tried in python3 and still having similar error.

Hi guys, first of all I wanted to thank you for this awesome framework!

I am curious if there is a way of adding new actions for the agents, i.e. a second type of attack which has different properties and effects.

I am learning how to use the magent toolkit. Below, i list some questions:

1. It seems that PyTorch is much more friendly and efficient than tensorflow, especially in research community. does the builtin model support pytorch?

2. How to run train_battle_game.py in my own pc without gpu despite it's expected it's time consuming to train large battle game? change map_size from 125 to 16? Actually, I only want to learn how to use this key APIs and embed this toolkit with my own project. Does any tutorial to introduce the key APIs?

Hi,

After I install MAgent, I try to run api_demo.py in examples, it seems that this example cannot find the directory build/render.

BTW, I run this program in PyCharm and set MAgent as the source folder.

Is anything wrong?

Best

zx@zx-MS-7A38:~/project/zx/MAgent$ python examples/train_pursuit.py
Traceback (most recent call last):
File "examples/train_pursuit.py", line 133, in
target_update=1000, train_freq=4
File "/home/zx/project/zx/MAgent/python/magent/model.py", line 153, in init
listener = multiprocessing.connection.Listener(addr)
File "/home/zx/anaconda2/lib/python2.7/multiprocessing/connection.py", line 132, in init
self._listener = SocketListener(address, family, backlog)
File "/home/zx/anaconda2/lib/python2.7/multiprocessing/connection.py", line 256, in init
self._socket.bind(address)
File "/home/zx/anaconda2/lib/python2.7/socket.py", line 228, in meth
return getattr(self._sock,name)(*args)
socket.error: [Errno 98] Address already in use
2017-12-12 21:01:39.301346: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
Process Process-1:
Traceback (most recent call last):
File "/home/zx/anaconda2/lib/python2.7/multiprocessing/process.py", line 267, in _bootstrap
self.run()
File "/home/zx/anaconda2/lib/python2.7/multiprocessing/process.py", line 114, in run
self._target(*self._args, **self._kwargs)
File "/home/zx/project/zx/MAgent/python/magent/model.py", line 297, in model_client
conn = multiprocessing.connection.Client(addr)
File "/home/zx/anaconda2/lib/python2.7/multiprocessing/connection.py", line 169, in Client
c = SocketClient(address)
File "/home/zx/anaconda2/lib/python2.7/multiprocessing/connection.py", line 308, in SocketClient
s.connect(address)
File "/home/zx/anaconda2/lib/python2.7/socket.py", line 228, in meth
return getattr(self._sock,name)(*args)
error: [Errno 111] Connection refused

env.render() doesn't do anything. Am I missing something?

bash build.sh
-- The C compiler identification is GNU 4.8.5
-- The CXX compiler identification is GNU 4.8.5
-- Check for working C compiler: /home/laker/anaconda3/bin/cc
-- Check for working C compiler: /home/laker/anaconda3/bin/cc -- works
-- Detecting C compiler ABI info
-- Checking if C linker supports --verbose
-- Checking if C linker supports --verbose - yes
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /home/laker/anaconda3/bin/c++
-- Check for working CXX compiler: /home/laker/anaconda3/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Checking if CXX linker supports --verbose
-- Checking if CXX linker supports --verbose - yes
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Configuring done
-- Generating done
-- Build files have been written to: /home/laker/MAgent/build
Scanning dependencies of target testlib
Scanning dependencies of target render
Scanning dependencies of target magent
[ 6%] Building CXX object CMakeFiles/render.dir/src/render/backend/render.cc.o
[ 9%] Building CXX object CMakeFiles/render.dir/src/render/backend/text.cc.o
[ 6%] Building CXX object CMakeFiles/render.dir/src/render/backend/data.cc.o
[ 12%] Building CXX object CMakeFiles/render.dir/src/render/backend/websocket.cc.o
[ 15%] Building CXX object CMakeFiles/render.dir/src/render/backend/utility/config.cc.o
[ 18%] Building CXX object CMakeFiles/testlib.dir/src/runtime_api.cc.o
[ 21%] Building CXX object CMakeFiles/render.dir/src/render/backend/utility/logger.cc.o
[ 24%] Building CXX object CMakeFiles/magent.dir/src/runtime_api.cc.o
[ 27%] Building CXX object CMakeFiles/magent.dir/src/temp_c_booster.cc.o
/home/laker/MAgent/src/render/backend/utility/logger.cc: In member function ‘void magent::render::Logger::log(const string&)’:
/home/laker/MAgent/src/render/backend/utility/logger.cc:14:20: error: ‘put_time’ is not a member of ‘std’
ostream << std::put_time(std::localtime(&time), "[%Y-%m-%d %H:%M:%S] [")
^
CMakeFiles/render.dir/build.make:182: recipe for target 'CMakeFiles/render.dir/src/render/backend/utility/logger.cc.o' failed
make[2]: *** [CMakeFiles/render.dir/src/render/backend/utility/logger.cc.o] Error 1
make[2]: *** Waiting for unfinished jobs....
[ 30%] Building CXX object CMakeFiles/magent.dir/src/gridworld/AgentType.cc.o
[ 33%] Building CXX object CMakeFiles/testlib.dir/src/temp_c_booster.cc.o
[ 36%] Building CXX object CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o
/home/laker/MAgent/src/gridworld/GridWorld.cc:456:0: warning: ignoring #pragma omp declare [-Wunknown-pragmas]
#pragma omp declare reduction (merge : std::vector : omp_out.insert(omp_out.end(), omp_in.begin(), omp_in.end()))
^
[ 39%] Building CXX object CMakeFiles/magent.dir/src/gridworld/Map.cc.o
[ 42%] Building CXX object CMakeFiles/magent.dir/src/gridworld/RenderGenerator.cc.o
[ 45%] Building CXX object CMakeFiles/magent.dir/src/gridworld/RewardEngine.cc.o
/home/laker/MAgent/src/gridworld/GridWorld.cc: In member function ‘virtual void magent::gridworld::GridWorld::step(int*)’:
/home/laker/MAgent/src/gridworld/GridWorld.cc:474:40: error: expected ‘+’, ‘’, ‘-’, ‘&’, ‘^’, ‘|’, ‘&&’, ‘||’, ‘min’ or ‘max’ before ‘merge’
#pragma omp parallel for reduction(merge: render_attack_buffer)
^
CMakeFiles/magent.dir/build.make:134: recipe for target 'CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o' failed
make[2]: *** [CMakeFiles/magent.dir/src/gridworld/GridWorld.cc.o] Error 1
make[2]: *** Waiting for unfinished jobs....
[ 48%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/AgentType.cc.o
[ 51%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/GridWorld.cc.o
[ 54%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/Map.cc.o
/home/laker/MAgent/src/gridworld/GridWorld.cc:456:0: warning: ignoring #pragma omp declare [-Wunknown-pragmas]
#pragma omp declare reduction (merge : std::vector : omp_out.insert(omp_out.end(), omp_in.begin(), omp_in.end()))
^
[ 57%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/RenderGenerator.cc.o
/home/laker/MAgent/src/gridworld/GridWorld.cc: In member function ‘virtual void magent::gridworld::GridWorld::step(int)’:
/home/laker/MAgent/src/gridworld/GridWorld.cc:474:40: error: expected ‘+’, ‘*’, ‘-’, ‘&’, ‘^’, ‘|’, ‘&&’, ‘||’, ‘min’ or ‘max’ before ‘merge’
#pragma omp parallel for reduction(merge: render_attack_buffer)
^
[ 60%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/RewardEngine.cc.o
[ 63%] Building CXX object CMakeFiles/testlib.dir/src/gridworld/test.cc.o
CMakeFiles/testlib.dir/build.make:134: recipe for target 'CMakeFiles/testlib.dir/src/gridworld/GridWorld.cc.o' failed
make[2]: *** [CMakeFiles/testlib.dir/src/gridworld/GridWorld.cc.o] Error 1
make[2]: *** Waiting for unfinished jobs....
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/magent.dir/all' failed
make[1]: *** [CMakeFiles/magent.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
CMakeFiles/Makefile2:104: recipe for target 'CMakeFiles/testlib.dir/all' failed
make[1]: *** [CMakeFiles/testlib.dir/all] Error 2
CMakeFiles/Makefile2:141: recipe for target 'CMakeFiles/render.dir/all' failed
make[1]: *** [CMakeFiles/render.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2

I found the feature provided in the pursuit game only contains the previous action and reward. The embeding id and relative position is missing. How I may locate the agents since then?

How can i get the position of a specific agent in gridworld ?

When I typed in config.json and video_100.txt, it says
magent::render::RenderException "Invalid handle of the map data file"
the file video_100.txt is in the directory build/render
Any ideas what may the problem be?

Actions are discrete actions. They can be move, attack and turn.
The move and attack action can be comprehend easily.
However, what does turn means?
I check the frames in render. I found the turn action correspond half-line. So I am confused with the turn action.
Hope for your reply.

I don't understand what is the main difference, because the stored episodes are still in the ReplayBuffer, it take the same amount of memory ?

I consider the information provided in the doc about the observation of agents may be not sufficient enough. I wander what the group 2 or group 2 minimap stands for in the spatial observation. Besides, how many numbers the embedding ID in the non-spatial observation takes up in the sequence（【ID embedding, last action, last reward and normalized position】)？ I consider these will be helpful for me to better understand the game. Thank you.

Of course I can use something like selenium and save screenshots to form a video later.
But is there any easier or better way of doing this?

I used Ubuntu 14.02 and Anaconda with Python 3.5.2 to run this project. I downloaded this repository and run these commands:

cd MAgent
sudo apt-get install cmake libboost-system-dev libjsoncpp-dev libwebsocketpp-dev
bash build.sh
export PYTHONPATH=$(pwd)/python:$PYTHONPATH

But there raised some errors and I couldn't install the python package successfully. So I copied the file "Magent" to my python project and used pycharm to mark the "Python" file in it as the Sources Root. When I run the file "api_demo.py", it showed some error massages:

/home/sean/anaconda3/bin/python3.6 /home/sean/DRL/MAgent/examples/train_pursuit.py
Traceback (most recent call last):
File "/home/sean/DRL/MAgent/examples/train_pursuit.py", line 11, in
import magent
File "/home/sean/DRL/MAgent/python/magent/init.py", line 2, in
from . import utility
File "/home/sean/DRL/MAgent/python/magent/utility.py", line 12, in
from magent.builtin.rule_model import RandomActor
File "/home/sean/DRL/MAgent/python/magent/builtin/rule_model/init.py", line 2, in
from .rush import RushPredator
File "/home/sean/DRL/MAgent/python/magent/builtin/rule_model/rush.py", line 7, in
from magent.c_lib import _LIB, as_int32_c_array, as_float_c_array
File "/home/sean/DRL/MAgent/python/magent/c_lib.py", line 42, in
_LIB = _load_lib()
File "/home/sean/DRL/MAgent/python/magent/c_lib.py", line 21, in _load_lib
lib = ctypes.CDLL(path_to_so_file, ctypes.RTLD_GLOBAL)
File "/home/sean/anaconda3/lib/python3.6/ctypes/init.py", line 348, in init
self._handle = _dlopen(self._name, mode)
OSError: /home/sean/DRL/MAgent/python/magent/../../build/libmagent.so: cannot open shared object file: No such file or directory

Process finished with exit code 1

It seems that my file miss some resources. But I use files completely copied from git. Maybe I missed some steps. I need some help to use the repository. o_o

Is there any support for environments where agents cannot go through each other?

For example,
in the following scenario, even if agent a has a moving speed of 2, it still cannot move away.

 b
bab
 b

是不是每次开机后再运行都需要重新安装或source一下？因为第一个例子跑过之后，第二个就会提示无法找到“”magent”

examples/show_arrange.py can not run in Python3.5 and Python3.6 on Ubuntu 16.04:
Traceback (most recent call last):
File "examples/show_arrange.py", line 22, in
PyGameRenderer().start(Server(messages=args.mess, mode=args.mode), grid_size=3.5)
File "/data/Work/MAgent/python/magent/renderer/server/arrange_server.py", line 345, in init
generate_map(mode, env, map_size, food_handle, handles, messages, font)
File "/data/Work/MAgent/python/magent/renderer/server/arrange_server.py", line 282, in generate_map
draw(base_x, base_y, scale, data)
File "/data/Work/MAgent/python/magent/renderer/server/arrange_server.py", line 270, in draw
for x in range(start_x, start_x + scale):
TypeError: 'float' object cannot be interpreted as an integer.

Python2.7 is OK.

I got a MemroyError when I run the example script train_pursuit.py. The traceback is copied as follow,

Process Process-1:
Traceback (most recent call last):
  File "/home/dl/.pyenv/versions/2.7.12/lib/python2.7/multiprocessing/process.py", line 258, in _bootstrap
    self.run()
  File "/home/dl/.pyenv/versions/2.7.12/lib/python2.7/multiprocessing/process.py", line 114, in run
    self._target(*self._args, **self._kwargs)
  File "/home/dl/Downloads/MAgent-master/python/magent/model.py", line 294, in model_client
    model = RLModel(**model_args)
  File "/home/dl/Downloads/MAgent-master/python/magent/builtin/tf_model/dqn.py", line 143, in __init__
    self.replay_buf_view     = ReplayBuffer(shape=(memory_size,) + self.view_space)
  File "/home/dl/Downloads/MAgent-master/python/magent/builtin/common.py", line 9, in __init__
    self.buffer = np.empty(shape=shape, dtype=dtype)
MemoryError

Actions in MAgent are defined as a number like 0 to 13, but how can we know what is the 0 action means? move left or right?

I was getting this error even though I installed boost library properly through both brew and macports.

| => bash build.sh 
-- Configuring done
-- Generating done
-- Build files have been written to: /Users/donghun/cs/MAgent/build
[ 75%] Built target testlib
[ 78%] Built target magent
[ 81%] Linking CXX executable render/render
ld: library not found for -lboost_system
clang-5.0: error: linker command failed with exit code 1 (use -v to see invocation)
make[2]: *** [render/render] Error 1
make[1]: *** [CMakeFiles/render.dir/all] Error 2
make: *** [all] Error 2

First, check if you properly installed boost
ls /usr/local/lib | grep --color boost_system
results:

libboost_system-mt.a
libboost_system-mt.dylib

Fix:
Change boost_system library to boost_system-mt (multi thread)

vi build/CMakeFiles/render.dir/link.txt
change boost_system to boost_system-mt in 2 places

and the build works.

Hello, I made new MAgent example.
I added cure function and made Support Agent like Medic in StarCraft.
And I edited some code to make Support Agent is ally with normal Agent.

Is there any rules I have to know about the pull request??

I work on https://github.com/arcane1028/MAgent/tree/dev-support
and result like this https://youtu.be/8RXB1f_uT7k

Thanks for your brilliant code.

I want to know how can I use the RL algorithm in my onw task? For example , If I have my own environment to provide reward and many agents, how can I use your RL code to train on my dataset? In other words, I just want to know what is 'parameter-sharing DQN' and which part of the code I can use to train my own dataset?

Hi , is it possible to compile / install MAgent on Windows 10?

Best regard / Rock

Who gets the kill_reward when two agents attacked the same grid?

I have run the code (train battle) and watch the video. However, I found the agents have learned nothing. Why this happens?Do I need to prolong the training process?

ModuleNotFoundError: No module named 'magent'

no module named 'magent' how do i fix this?

run show_battle_game.py error

Traceback (most recent call last):
  File "examples/show_battle_game.py", line 9, in <module>
    from magent.renderer import PyGameRenderer
  File "/home/minghigh/Desktop/MAgent/python/magent/renderer/__init__.py", line 1, in <module>
    from base_renderer import BaseRenderer
ImportError: No module named 'base_renderer'

my PYTHONPATH is okey.

$python examples/train_pursuit.py --train
Traceback (most recent call last):
File "examples/train_pursuit.py", line 11, in
import magent
ModuleNotFoundError: No module named 'magent'

laha@laha-unbutun:$ uname -a
Linux laha-unbutun 4.18.0-12-generic #13-Ubuntu SMP Wed Nov 14 15:17:05 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
laha@laha-unbutun:$ g++ --version
g++ (Ubuntu 8.2.0-7ubuntu1) 8.2.0
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

laha@laha-unbutun:~$ cmake --version
cmake version 3.12.1

CMake suite maintained and supported by Kitware (kitware.com/cmake).

Thanks for your brilliant code.
When I open index.html in browser.The page show: magent : :render::RenderException: "invalid handle of the map configuration file",and didn't show videos.I want to know the reason.

File "examples/show_battle_game.py", line 9, in
from magent.renderer import PyGameRenderer
File "/Users/lanjiedediannao/MAgent/python/magent/renderer/init.py", line 2, in
from .pygame_renderer import PyGameRenderer
File "/Users/lanjiedediannao/MAgent/python/magent/renderer/pygame_renderer.py", line 6, in
import pygame
ImportError: No module named pygame
安装的过程没有报错，可是在执行show_battle_game的时候出现了这个错误，希望得到解答

Spatial view consists of several rectangular channels, Walls, group 1, group 1's hp, group 1's minimap, group 2, group 2's hp, group 2's minimap. I have following questions:

1. What's the meaning of hp?
2. What are included in the minimap? It contains the walls?
3. All these values should be o or 1. However, when I print some observation (Spatial view) out, it has other values like 1.667. I am very confused.

I really appreciate that if you can replay to it at your appropriate time.

(I already run show_battle_game.py successfully with python2)

python3 examples/show_battle_game.py

error as follows

Traceback (most recent call last):
  File "examples/show_battle_game.py", line 15, in <module>
    PyGameRenderer().start(Server())
  File "/home/minghigh/Desktop/MAgent/python/magent/renderer/pygame_renderer.py", line 281, in start
    now_group[1] * grid_size
  File "/home/minghigh/Desktop/MAgent/python/magent/renderer/pygame_renderer.py", line 52, in draw_rect
    map(int, (round(a[0]), round(a[1]), round(w + a[0] - round(a[0])), round(h + a[1] - round(a[1]))))
TypeError: Argument must be rect style object

After running train_multi.py, I cannot render the environment.
After typing config.json and video_*.txt generated by the command
python3 examples/train_multi.py into the input boxes,
It says invalid map file data.

python examples/train_pursuit.py --train

libgomp: Invalid value for environment variable OMP_NUM_THREADS
Traceback (most recent call last):
File "examples/train_pursuit.py", line 133, in
target_update=1000, train_freq=4
File "/home/hello/magent/python/magent/model.py", line 153, in init
listener = multiprocessing.connection.Listener(addr)
File "/usr/lib/python2.7/multiprocessing/connection.py", line 132, in init
self._listener = SocketListener(address, family, backlog)
File "/usr/lib/python2.7/multiprocessing/connection.py", line 256, in init
self._socket.bind(address)
File "/usr/lib/python2.7/socket.py", line 228, in meth
return getattr(self._sock,name)(*args)
socket.error: [Errno 98] Address already in use
Process Process-1:
Traceback (most recent call last):
File "/usr/lib/python2.7/multiprocessing/process.py", line 258, in _bootstrap
self.run()
File "/usr/lib/python2.7/multiprocessing/process.py", line 114, in run
self._target(*self._args, **self._kwargs)
File "/home/hello/magent/python/magent/model.py", line 294, in model_client
model = RLModel(**model_args)
File "/home/hello/magent/python/magent/builtin/tf_model/dqn.py", line 94, in init
self.qvalues = self._create_network(self.input_view, self.input_feature, use_conv)
File "/home/hello/magent/python/magent/builtin/tf_model/dqn.py", line 164, in _create_network
h_conv1 = tf.layers.conv2d(input_view, filters=kernel_num[0], kernel_size=3,
AttributeError: 'module' object has no attribute 'layers'