questions

myriad.md

Initialization of a Binary:

• Calculate dt_{b_{i}} according to Eq. 22

• Calculate the \eta_b using the condition: (Page 8, right column, second paragaph)

• ETA_B = D_TIME_MIN / (2.0 * min(dt_{b_{i}});

• Initialization of dt = D_TIME_MIN / 2 (Question1: This for everyone? or we just use the minimum?)

Integration of a Binary:

• For each binary, we follow the procedure:

• Determinate new dt, calculating the minimum of the members (Question 2: we already initialized one, so they are the same?)

• (Huge Question 3)

• I understand that for a certain t we determinate if it's even or odd according to the dt, and what to do in both cases but what is the meaning of the phrase "If t is odd, we maintain the same time step and by using DELTA t_old we determine the time step at both the beginning dt_{0} and the end dt_{1} of that time step and find their average: dt = (dt_0 + dt_1)/2"

• Let's do an example:

  • We have a certain time tt and our timestep dtt, at the end, the time is odd, so we maintain the dtt (DELTA t_old)
    • What is dt0?
    • What is dt1?
    • If dt0 = dtt ,our dt1 will also be dtt?
      • is that is the case, dt = (dt0 + dt1)/2 = (dtt + dtt)/2 = dtt
      • So the restriction makes no sense because dt/2 < dtt = dt < dt.

• Prediction (Get Predicted pos/vel)

• 3 times

• Update forces using predicted positions and velocities

• Calculate perturbation forces (Question 4: This implementation is based on which work?, See lasts comments of your email)

• correct positions and velocities (Question 5: Normal Hermite, updates the REAL pos/vel, but since here we are repeating the force update and correction process, we correct the PREDICTED pos/vel ?)

---

About your email

We find new time steps for the particles. If a binary is not terminated, its corresponding ghost does not change time step.

There is another special treatment for the ghost particles? I Predict-Update Forces-Correct positions/velocities normally as the other particles, but only this conditions of the timesteps is required?

Yes, only over the neighbours. I trust the list of neighours because I have spent a lot of time on it, so there is no need to search among all particles for perturbers. To be on the safe side, you can have the neighbour radius a little large, so you don't miss anything. I have to say here that I don't tryst very much Myriad's treatment of perturbers. It is working, but I did not test it very much. It is hard to test it anyway.

Let's assume the next steps

Integration (using int times t=0,1,2,3,4,5...)

1. If there are binaries -> integration
2. Prediction (N)
3. Update forces (Nact) <- calculate neighbor list
4. Check for new binaries
5. Correction (N)

• At a t=0 I found binaries on step 4,
• so the next time (loop) t=1 I enter into the step 1,
• but since I formed the ghost particle in the previous step
• I don't have the neighbor list which is calculated in the step 3
• Should I build the neighbor list for every ghost particle when it's created?
• Or for the "first" evolution of the binary I use one of the members neighbor's list?