Sequential optimization and constraints

Sequential optimization can be used to apply soft constraints. When you select a sequential Objective such as "Flow", this is now considered an objective rather than a hard constraint. The solver will internally penalize values that violate the constraint by applying a cost without rendering the model infeasible. For example, assume your model has a flow constraint with a Max = 100. When you define a sequential Objective for "Flow", any flow less than or equal to 100 incurs no cost penalty. Flow above 100 incurs a cost that the solver will try to minimize.

The use of these soft constraints is extremely valuable in analyzing your network. You can determine how much of an element within your network is actually required. Sequential optimization will minimize the overage, letting you identify the answers to questions of how big, how many, etc. For example, if you define your work centers so that the time capacity is 0, then run a sequential optimization using "Work Center Throughput Capacity", the solver determines the lowest actual number of hours required.

In cases where your model is infeasible and you have no constraints, when you define a sequential objective, this model will still be infeasible when running sequential optimization. In cases where the model was feasible, the priority 2 problem could solve, but would use the priority 2 objective(s) as soft constraints.

Last modified: Friday May 12, 2023

Is this useful?