diff --git a/README.md b/README.md
index 36a9393c4600ef08fe46722e4023657336875983..66cc5000948c3a2a009ee9fcbc2a9bbadb9eabdc 100644
--- a/README.md
+++ b/README.md
@@ -57,59 +57,52 @@ print("Score: ", np.sum(reward)) # Print the total reward accumulated
This is the docstring for the `stirap-v0` environment:
- Description:
- A potential with three wells is used to confine a quantum system:
+ Description:
+ A potential with three wells is used to confine a quantum system:
- v(x) = 0.5 * self.trap_strength * (x - xl)**2 * x ** 2 * (x - xr)**2
+ v(x) = 0.5 * self.trap_strength * (x - xl)**2 * x ** 2 * (x - xr)**2
- The dynamics is described by the 1D Schrödinger equation.
+ The dynamics is described by the 1D Schrödinger equation.
- The system is initially in the ground state of the left well. The goal is to move as much of the
- probability density to the right well by the end of the dynamics.
-
- The agent can move the left and right wells left independently by an amount Delta at each timestep.
-
- Source:
-
- Observation:
-
- If full_observation=False:
- Type: Box(4)
- Num Observation Min Max
- 0 Left population 0 1
- 1 Right population 0 1
- 2 Left well position -2 +2
- 3 Right well position -2 +2
-
- If full_observation=True:
- Type: Box(2 * n + 2)
- where n is the number of space points.
-
- (re(psi), im(psi), left_well_pos, right_well_pos)
-
- Actions:
- Type: Discrete(9)
- Num Action [ Left, Right ]
- 0 [-Delta, -Delta],
- 1 [ 0. , -Delta],
- 2 [ Delta, -Delta],
- 3 [-Delta, 0. ],
- 4 [ 0. , 0. ],
- 5 [ Delta, 0. ],
- 6 [-Delta, Delta],
- 7 [ 0. , Delta],
- 8 [ Delta, Delta]
-
- Reward:
- Reward at each time step is proportional to the population in the right well times t
- Reward -10 is given if the episode terminates before hand
-
- Starting State:
- The system is initially in the ground state of the left well.
-
- Episode Termination:
- * After the amount of time defined by env.totaltime (and the timesteps defined by env.timesteps)
- * If the agent moves the traps out of range (or swaps them)
+ The system is initially in the ground state of the left well. The goal is to move as much of the
+ probability density to the right well by the end of the dynamics.
- Solved Requirements:
- Handmade solution scores 120. Try and beat it
\ No newline at end of file
+ The agent can move the left and right wells left independently by an amount Delta at each timestep.
+
+ Source:
+
+ Observation:
+
+ If full_observation=False:
+ Type: Box(4)
+ Num Observation Min Max
+ 0 Left population 0 1
+ 1 Right population 0 1
+ 2 Left well position -2 +2
+ 3 Right well position -2 +2
+
+ If full_observation=True:
+ Type: Box(2 * n + 2)
+ where n is the number of space points.
+
+ (re(psi), im(psi), left_well_pos, right_well_pos)
+
+ Actions:
+ Type: Box(2)
+ Num Action Min Max
+ 0 Move left well of amount -1 +1
+ 1 Move right well of amount -1 +1
+
+ Reward:
+ Reward at each time step is proportional to the population in the right well times t
+ Reward -10 is given if the episode terminates before hand
+
+ Starting State:
+ The system is initially in the ground state of the left well.
+
+ Episode Termination:
+ * After the amount of time defined by env.totaltime (and the timesteps defined by env.timesteps)
+ * If the agent moves the traps out of range (or swaps them)
+
+ Solved Requirements:
+ Handmade solution scores 120. Try and beat it
\ No newline at end of file