In some cases, the only limits to the decisions you can take are the bounds on the decision variables themselves. In the toy manufacturing example above, you would obviously not want to be recommended that a negative number of some toy be produced. Such rules can be modeled simply by setting bounds for decision variables when they are being declared.
In most applications, there will be side constraints that limit the options of what is a legal set of decisions. In the pick-up and delivery example above, you will want to specify that each truck only holds a certain capacity. Restrictions like this are modeled by adding constraints to your problem. Constraints can take many different forms. One very common type of constraint says that a certain sum a quantities cannot exceed a certain amount. In our Seeker documentation, we introduce you to the constraints you can impose to let the computer know what set of decisions will actually work in practice.