There seems to be an issue with accessing the solution of a 2-stage stochastic programming model after using PG with Asynchronous option to solve the same. The error occurs when we use JuMP.value(model[2, :x], 25) for say a 25 scenario problem. The issue arises when I try to access the solution of the decision variables in the last scenario. This error is erratic, some times it throws an error with a certain number of scenarios and it does not with others. The error does not occur when we do the same solve in the Serial mode.
There seems to be an issue with accessing the solution of a 2-stage stochastic programming model after using PG with Asynchronous option to solve the same. The error occurs when we use
JuMP.value(model[2, :x], 25)
for say a 25 scenario problem. The issue arises when I try to access the solution of the decision variables in the last scenario. This error is erratic, some times it throws an error with a certain number of scenarios and it does not with others. The error does not occur when we do the same solve in the Serial mode.Here is the stack trace: