Closed etatara closed 11 months ago
It seems like just using JuMP.value(node[:x]) works OK.
Thanks for reporting this. Both versions are supposed to work. I'll push out a fix asap.
This should now be fixed by #93.
Just FYI, fixing this revealed another issue with running optimize!
on a node after a graph. It is normally possible to track separate solutions on a node versus the graphs that contain it. Right now, if you optimize a node directly, it will clear all of solutions on each graph. I'll plan to file an issue with a working example.
JuMP.value(node::OptiNode,vref::VariableRef) throws an error:
This can be replicated in the example optigraph_4_nodes.jl by calling:
In optinode.jl:48, the call to JuMP.backend(node).result_location[node.id] throws the error because the node.id symbol is different from the key in the result_location dict.