facebookexperimental / facebookexperimental/Robyn
Budget allocator - Bounded spend is not matching with initial or new given spend in the input
- Dominant language
- Jupyter Notebook
- Stars
- 1.5k
- Forks
- 433
- PR merge metrics
- No merged PRs in 30d
Description
## Project Robyn
## Describe issue
I wanted to run two scenarios - 1) base scenario with NULL spend i.e. 6.3MM initial spend and 2) new budget scenario with initial+$971,510 for a date range date_range = c("2023-08-03", "2023-09-02"). When I ran scenario 2, I am getting bounded spend which is only $100K higher than initial spend and unbounded spend is coming out correctly as desired i.e. initial + $971,510. I don't understand why bounded spend is so random?
## Provide reproducible example

`AllocatorCollect <- robyn_allocator(
#InputCollect = InputCollect,
#OutputCollect = OutputCollect,
json_file = json_file,
dt_input = input_data,
dt_holidays = dt_prophet_holidays,
select_model = "4_61_27",
date_range = c("2023-08-03", "2023-09-02"), # Default last month as initial period
total_budget = 7290763, # When NULL, default is total spend in date_range
channel_constr_low = c(1, 1, 1, 1, 1, 1, 0.8, 0.8, 0.8, 0.8,0.8, 0.8, 0.8, 0.8, 0.8,0.8, 0.8, 0.8, 0.8, 0.8,0.8, 0.8, 0.8, 0.8, 0.8,0.8, 0.8, 0.8, 0.8),
channel_constr_up = c(1, 1, 1, 1, 1, 1, 1.2, 1.2, 1.2, 1.2,1.2, 1.2, 1.2, 1.2, 1.2,1.2, 1.2, 1.2, 1.2, 1.2,1.2, 1.2, 1.2, 1.2, 1.2,1.2, 1.2, 1.2, 1.2),
channel_constr_multiplier = 3,
scenario = "max_response",
export = TRUE
)`
## Environment & Robyn version
Robyn version 3.10
Contributor guide
Assessment
This issue has not been assessed yet.