I have 15 treatments. My main interest is to find the best treatment. The response is measured every day up to 30 days. My model has an interaction effect between time and treatments. I use suitable effect size, but I need power of 80% with type I error 5%. How can I calculate sample size by simulation?