Community Profile

photo

Emmanouil Tzorakoleftherakis


Last seen: 1 day ago

MathWorks

136 total contributions since 2018

Emmanouil Tzorakoleftherakis's Badges

  • Knowledgeable Level 3
  • 6 Month Streak
  • Revival Level 2
  • First Answer

View details...

Contributions in
View by

Answered
Confusion in Critic network architecture design in DDPG
Hello, Does this paper use DDPG as well? Any images that show the network architecture? If it's another algorithm, the critic m...

2 days ago | 0

Answered
Deep reinforcement learning for multi-agents
Hello, The policies you will get from RL training change depending on the amount of time the agents spend exploring. Usually, i...

5 days ago | 1

| accepted

Answered
How to implement multi-agent RL training with custom MATLAB environment function?
Hello, Currently multi-agent training is only supported in Simulink. If you have an environment created in MATLAB you could cop...

5 days ago | 1

| accepted

Answered
Reinforcement learning DDPG action fluctuations
Hello, During training, DDPG explores the action space by adding noise to the output of the actor (see step 1 here). That expla...

5 days ago | 0

| accepted

Answered
Import patternnet into deep Q network
Hello, By 'patternnet' are you referring to this function? If yes, then this does not use the deep learning layer objects that ...

11 days ago | 0

| accepted

Answered
how to change the reward parameter in RL toolbox every time step?
One way is to implement your logic in your reward subsystem using a MATLAB Fcn block. Execute the MATLAB Fcn block every Ts time...

11 days ago | 0

| accepted

Answered
Training a neural network in a real world system (inverted pendulum)
Hello, I am not sure why you get zero values consistently, but since you say this works when you don't have the agent block in ...

11 days ago | 0

| accepted

Answered
Invalid input argument type or size such as observation, reward, isdone or loggedSignals. (Reinforcement learning toolbox)
Hello, It's challenging to reproduce this without having access to a reproduction model (including the environment definition)....

14 days ago | 0

| accepted

Answered
Reinforcement learning actions using DDPG
Hi Jason, In the documentation link you provided it's mentioned "Variance*sqrt(Ts) be between 1% and 10% of your action range"....

16 days ago | 0

| accepted

Answered
DDPG Not getting any reward? Using the example from MathWorks
Hi Jun, The IsDone input to the agent block is always true, so all episodes end prematurely. This is why you don't see any chan...

16 days ago | 0

Answered
Timestep problems using Reinforcement Learning with EnergyPlus
Hello, I am not familiar with EnergyPlus so I am not sure about the error you are getting (although it seems like you are tryin...

17 days ago | 0

Answered
The reward is reduced after certain episode on the DQN
Hello, There is no guarantee that the reward will keep going up always when using RL (after all there is certain stochastisity ...

17 days ago | 0

Answered
Error on Specifying observation using rlFiniteSetSpec(Reinforcement Learning)
Hello, The way you called rlFiniteSetSpec basically says that you have two observations and that they are both equal to 1. The ...

17 days ago | 1

| accepted

Answered
How to log/record signal data from a reinforcement learning (RL) agent/environment over multiple training episodes using Simulink?
If you call 'train' with an output argument, you will get many useful statistics from training (see here). Would this work?

1 month ago | 1

| accepted

Answered
How to make simulink function block interact with RL Agents by updating states and other environment parameters?
I think you can just read the workspace params if you set an additional variable as a parameter in the MATLAB Fcn block. In the ...

1 month ago | 0

Answered
invalid input argument during training of custom rl-environment
It seems this is a problem with how you step function populates the states and outputs, or maybe you need to transpose somewhere...

1 month ago | 0

Answered
What is the best activation function to get action between 0 and 1 in DDPG network?
Hello, With DDPG, a common thing to do in the final 3 layers of the actor is to use a fully connected layer, a tanh layer and a...

1 month ago | 0

| accepted

Answered
Expected reward blows up while training (DDPG agent, reinforcement learning)
Hello, This answer may be helpful. I would make sure your reward signal outputs values that make sense, and also possibly simp...

2 months ago | 0

Answered
parts of Walking Robots from Sebastian Castro
Hello, Is the video you are looking for listed here perhaps?

2 months ago | 0

Answered
Dynamically change a block input during Deep Reinforcement Learning after each episode. How To?
Hello, Please have a look at this example, and in particular the 'localResetFcn' at the bottom ('setBlockParameter'in the reset...

2 months ago | 0

Answered
Use saved reinforcement learning DDPG agent
Hello, Please see my response here. In short, the behavior you see during training and after training are not expexted to match...

2 months ago | 0

Answered
warning in reinforcement training
Seems the reset function has not been defined properly. Have a look at this example for some ideas.

2 months ago | 0

| accepted

Answered
Is it practicable to train multiple agents simutaneously using RL Toolbox?
As of R2020b release, Reinforcement Learning Toolbox lets you train multiple agents simultaneously in Simulink. Please see follo...

2 months ago | 0

Answered
Reinforcement learning toolbox multiple agents training
As of R2020b release, Reinforcement Learning Toolbox lets you train multiple agents simultaneously in Simulink. Please see follo...

2 months ago | 3

Answered
Is it possible to carry out multi agent reinforcement learning in MATLAB ?
As of R2020b release, Reinforcement Learning Toolbox lets you train multiple agents simultaneously in Simulink. Please see follo...

2 months ago | 0

Answered
some of the saved agents in DQN reinforcement learning algorithm do not reproduce the training rewards
This is a common misconception. When you train an agent and you get a certain reward out of a specific episode, if you were to s...

2 months ago | 0

| accepted

Answered
Deep Q-Network Rewards incorporation?
Hello, If you have a look at this page, it shows where the reward is incorporated in a custom MATLAB environment. As you can se...

2 months ago | 0

| accepted

Answered
reinforcement learning and DDPG agent problem
Looks like training was not successful. There could be many things at fault here - some suggestions: 1) Make sure you are rando...

2 months ago | 1

| accepted

Answered
RL agent error using simulink
Looks like when you call getActionInfo, instead of passing an environment or agent object as an argument, you are passing a doub...

2 months ago | 0

| accepted

Answered
Deep Q-network reinforcement learning
Hello, The functionality to customize the action space is not yet available. A couple of workarounds: 1) Use penalties in the ...

2 months ago | 0

| accepted

Load more