Use this tolerance to place an upper limit on the number of iterations the solver will perform. An iteration is the fundamental operation performed by the solver. At the risk of oversimplification, an iteration is a process that involves forcing a variable, currently at a zero value, to become nonzero until some other variable is driven to zero, improving the objective as we go. In general, larger models will take longer to perform an iteration, and nonlinear models will take longer than linear models. The default iteration limit is 0, meaning no limit is imposed on the iteration count.

If the solver hits this limit, it returns to normal command mode. If the model contains integer variables, LINGO will restore the best integer solution found so far. You may need to be patient, however, because the solver may have to perform a fair amount of work to reinstall the current best solution after it hits a runtime limit.

Note:Some caution is required when interrupting the solver.  There must be an incumbent solution available if you hope to interrupt the solver and have it return a valid solution.  You can always tell if an incumbent solution is available by examining the Best Obj field in the Extended Solver Status box of the solver status window.  If this field is blank, then an incumbent solution does not exist, and the solution returned after an interrupt will be invalid.  If, on the other hand, this field contains a numeric value, then you should be able to interrupt and return to a valid, if not globally optimal, solution.