1: \begin{abstract}

2:   Simulations of binary collisions involving compact objects require initial data

3:   that satisfy the constraint equations of general relativity.

4:   For binary boson star simulations it is common practice to use a superposition

5:   of two isolated star solutions to construct an approximate solution to the constraint equations.

6:   Such superposed data is simple to set up compared to solving these equations explicitly,

7:   but also introduces extra constraint violations in the time evolution.

8:   %

9:   %

10:   %

11:   In this work we investigate how physical observables depend on

12:   the quality of initial data in the case of head-on boson star collisions.

13:   In particular we compare results obtained from data prepared using

14:   four different methods: the standard method to superpose isolated stars,

15:   a heuristic improvement to this superposition technique

16:   and two versions of this data where excess constraint violations

17:   were removed through a conformal thin-sandwich solver.

18:   We find that differences in the time evolutions are dominated

19:   by differences in the way the two superposition methods differ,

20:   whereas additionally constraint solving the superposed data has smaller impact.

21:   The numerical experiments are conducted using the pseudospectral code \bamps{}.

22:   Our work demonstrates that \bamps{} is a code suited for generating high accuracy

23:   numerical waveforms for boson star collisions due to the exponential convergence in the polynomial resolution

24:   of the numerical approximation.

25: \end{abstract}

26: