From 29be64648e5b972713bba2bfe47ca579b1dab9d4 Mon Sep 17 00:00:00 2001
From: "S.P. Mohanty" <spmohanty91@gmail.com>
Date: Thu, 24 Oct 2019 18:42:06 +0200
Subject: [PATCH] Mark agents as done in dont dict on max episode steps
violation
---
flatland/envs/rail_env.py | 2 ++
1 file changed, 2 insertions(+)
diff --git a/flatland/envs/rail_env.py b/flatland/envs/rail_env.py
index 8cbab637..b5debcd4 100644
--- a/flatland/envs/rail_env.py
+++ b/flatland/envs/rail_env.py
@@ -488,6 +488,8 @@ class RailEnv(Environment):
self.rewards_dict = {i: self.global_reward for i in range(self.get_num_agents())}
if (self._max_episode_steps is not None) and (self._elapsed_steps >= self._max_episode_steps):
self.dones["__all__"] = True
+ for i_agent in range(self.get_num_agents()):
+ self.dones[i] = True
return self._get_observations(), self.rewards_dict, self.dones, info_dict
--
GitLab