How to discretize statespace "logarithmicly" arround 0?
1 vue (au cours des 30 derniers jours)
Afficher commentaires plus anciens
Hello, I am trying to discretize my statespace in a way such that values arround 0 are closer together and values further away from 0 are further appart from another. I tried to do it logarithmicly but that only works for values >0 and i need them to be negative as well.
Later in the programm i have to calculate a new state, with is most likely not on my grid. Therefor i have to find the closest neighbor.
My current solution looks like this:
% States in phi
phi_spaceing=logspace(-4,pi,10)
phi_spaceing_log=log10(phi_spaceing)
states_phi=cat(2,-flip(phi_spaceing),0,phi_spaceing)
% Find neares Neighbor in phi-Direction
if abs(state_vector_calculation(3,1)) <= (10^-4)/2
calculated_state_phi_index = 11;
elseif state_vector_calculation(3,1) > 10^-4
calculated_state_phi_index = 11 + find(abs(phi_spaceing_log-log10(state_vector_calculation(3,1)))<0.2498);
elseif state_vector_calculation(3,1) < -10^-4
calculated_state_phi_index = 11 - find(abs(phi_spaceing_log-log10(-1*state_vector_calculation(3,1)))<0.2498);
end
I was Wondering if there is any better solution to this problem?
Thank you in advance for your help.
0 commentaires
Réponses (1)
Himanshu
le 25 Avr 2024
Hey,
Your approach to discretize the state space using a logarithmic scale is reasonable, but it seems you're encountering some challenges, particularly with negative values.
Instead of logarithmic spacing, you might consider using a different transformation that can handle both positive and negative values effectively. One such transformation is the hyperbolic tangent (tanh) function.
The following code snippet gived a rough idea of how to make these changes in the code.
% Define the range for phi values
phi_min = -pi;
phi_max = pi;
% Define the number of discrete states
num_states = 21; % Adjust as needed
% Discretize the state space using hyperbolic tangent function
phi_values = tanh(linspace(phi_min, phi_max, num_states));
% Find the nearest neighbor in the phi-direction
[~, calculated_state_phi_index] = min(abs(phi_values - state_vector_calculation(3,1)));
Hope this helps!
0 commentaires
Voir également
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!