I've built a model of the Solar System by calculating Keplerian orbital elements. The model takes a time $t$ which is incremented with each frame of the simulation. I then calculate the new positions of each body. $t$ is really the time since perihelion for each orbit.
This means that at the beginning of the simulation each object starts at its perihelion (since $t=0$). So, now I'd like to place a new body (a comet, for example) with a position $P$ and velocity $V$. From $P$ and $V$, I calculate the orbital elements such as eccentricity $e$, semi-major axis $M$, etc.
Now, I can calculate the next position of this new body (since $t$ has been incremented) using these orbital elements. However, since $t$ is the time since this new orbits perihelion, the calculated position will not be the next incremental position. One solution to this could be that I calculate the time ($t\text'$) at which this new orbit would position the body at $P$ and increment $t\text'$. This means that instead of having a global $t$, I have a $t$ for each body which is incremented at the same rate.
Does this sound sensible?