Memory consumption growth with many large MILP's in JuMP

Specific Domains Optimization (Mathematical)
,
61

I am working on contacting FICO. They don’t make it easy.

2 Likes
62

v8.0.4 or v8.4?
8.0 is from 2016

63

Yikes yes 8.0. I didn’t realize that it is that old! I’ll make this priority one.

1 Like
64

Per discussion in Memory leaks · Issue #131 · jump-dev/Xpress.jl · GitHub I ran the following to check if postsolve would release more memory

using REoptLite, JuMP
import Xpress

for i in range(1,stop=1000)
    m = direct_model(Xpress.Optimizer(OUTPUTLOG = 0))
    r = run_reopt(m, "outage.json")
    finalize(backend(m))
    Xpress.postsolve(backend(m).inner)
    GC.gc()
end

The above code was run using Xpress 8.0 and 8.12 with the following result:

add_postsolve
add_postsolve800×600 45.7 KB
.

It appears that Xpress 8.12 solves the problems faster but with larger memory consumption.

1 Like