sparisi / mips Goto Github PK

For the 3-obj exact lqr,

feval
'mexIntegrate_LQR3_P1_utopia' cannot be identified.

Hi, I'm interested in your PMGA algorithm. But I have some questions when I read the code.

1. In the GaussianLinearFixedvarDiagmean.m, you define two functions to obtain the derivative of the logarithm of the policy and hessian matrix of the logarithm of the policy, respectively. How do you obtain the exact formulations? I dont find the derivation process in your paper (AAAI'15 and the extension), could you pls provide some refs on these?

2. In the params_lqr_mo.m (also in the Sec. 6.1 of your Journal of Artificial Intelligence Research (2016)), you give the formulation of \rho in different parameter types, I also wonder about the derivation or refs on this.

Look forward to your reply. Thanks!

Best regards,
Shutong

Hi Simone,

I find it is somewhat hard to understand the terms "volume of the manifold" and "Vol(T)" in PMGA algorithm. I think in the case of 2-dimensional object, the volume of the manifold may be the area of J_1 and J_2, right? After printing L, V, D_theta_J_i, and D_t_theta_i of each iteration, I find that D_theta_J_i varies irregularly and items of D_t_theta_i varies according to the updating rho. It is interesting that V fluctuates but increases overall, and hence L increases overall (because the indicator doesn't vary widely). I'm so curious about this. How to understand V (and Vol(T)) in 2-dimensional or 3-dimensional case? And why it shows a fluctuating and increasing trend?

Look forward to your reply!

Best regards,
Shutong

It says

出错 RUN_Manifold (第 11 行)
policy = policy.makeDeterministic; % Learn deterministic low-level policy