Revision history [back]

1 Yes. There are no guarantees on that and I'd actually assume that you will not get the same results. First, I think you'd obviously talk about simulation and not real robots. But even then and if the simulation is running without any noise we don't have a round based system, where data is passed on synchronously. For example, a single laser scan might not arrive in the same time window between runs leading to a slightly different costmap and thus behaviour. This is even excluding the fact that algorithms might do random sampling. This will affect everything happening after that as it depends on each other.

That being said. We are talking about __exactly__ the same. Depending on the stability of the algorithms and simulation you can assume similar values as you would on a real world system.

2 The settings/configuration if you pass them in will be the same. The actually observed behavior will be affected by the same things as 1.

1 Yes. There are no guarantees on that and I'd actually assume that you will not get the same results. First, I think you'd obviously talk about simulation and not real robots. But even then and if the simulation is running without any noise we don't have a round based system, where data is passed on synchronously. For example, a single laser scan might not arrive in the same time window between runs leading to a slightly different costmap and thus behaviour. This is even excluding the fact that algorithms might do random sampling. This will affect everything happening after that as it depends on each other.

That being said. We are talking about __exactly__ exactly the same. Depending on the stability of the algorithms and simulation you can assume similar values as you would on a real world system.

2 The settings/configuration if you pass them in will be the same. The actually observed behavior will be affected by the same things as 1.