# Column Gerneration Example in the documentation/hot start after adding variables to LP
**URL:** https://discourse.julialang.org/t/column-gerneration-example-in-the-documentation-hot-start-after-adding-variables-to-lp/69603
**Category:** Optimization (Mathematical)
**Tags:** question, jump
**Created:** [October 12, 2021, 3:54am UTC](https://discourse.julialang.org/t/column-gerneration-example-in-the-documentation-hot-start-after-adding-variables-to-lp/69603 "2021-10-12T03:54:30Z")
**Posts on this page:** 2
**Page:** 1
### Author:  [@dirk](https://discourse.julialang.org/u/dirk)
#### Post date: [October 12, 2021, 3:54am UTC](https://discourse.julialang.org/t/column-gerneration-example-in-the-documentation-hot-start-after-adding-variables-to-lp/69603/1 "2021-10-12T03:54:30Z")
Hi all,
I just started using Julia and JuMP. I went through the tutorials
([https://jump.dev/JuMP.jl/stable/tutorials/Mixed-integer%20linear%20programs/cutting\_stock\_column\_generation/](https://jump.dev/JuMP.jl/stable/tutorials/Mixed-integer%20linear%20programs/cutting_stock_column_generation/)) and in the column generation example and in general it is unclear to me if the problem gets re-solved from scratch whenever I add a new variable to the model or if the model “in the solver” is updated and there is a hot start from the current solution.
More general, if a JuMP model is changed after it was solved and I solve the changed version, is it solving from scratch or is there a hot start in the solver?
From the documentation I guess it’s only changing the model in the solver when using solver independent callbacks ([Solver-independent Callbacks · JuMP](https://jump.dev/JuMP.jl/stable/manual/callbacks/)).
Thanks a lot in advance!
---
### Author:  [@odow](https://discourse.julialang.org/u/odow)
#### Post date: [October 12, 2021, 4:40am UTC](https://discourse.julialang.org/t/column-gerneration-example-in-the-documentation-hot-start-after-adding-variables-to-lp/69603/2 "2021-10-12T04:40:31Z")
It depends on the solver. In most cases, JuMP maintains the model in memory, so it doesn’t rebuild every time.