m****d 发帖数: 331 | 1 When i call Cplex in GGAMS to solve a MIP problem, it seems that Cplex gives
me weired results. It first fixes the binary variables to minimize the
objective function, then solve the rest problem as a LP problem.
This gives me infeasible solutions. for example, a constraint like:
q(i,g,t)<=M*u(i,g,t), where q are contious variables and u are bianry
variables, M is just a parameter.
In the .list file, i found that Cplex just gives 0 to all u(i,g,t). But, q(i,g
,t) has to be positive so that it ca |
|