Tribium DECISIONS Blog

blog about how to make good decisions

How to use constraints

without comments

Constraints are knowns as the limitations of the problem space applied while evaluating all identified alternatives using specific decision model. Constraint is simply a condition which must be met in order to assume that given alternative is acceptable. If the constraint condition for given alternative has been not satisfied it usually means that this alternative did not pass the basic test and cannot be considered as a valid alternative and should be discarded.

It is possible that we might find an alternative with the greatest file you of the final a violation (utility)  in the set that are being evaluated, but having violated any of the defined constraints. The situation is in perfect concord with our intuition and experience. It might easily give some examples of decisions for which we had almost ideal best alternatives which unfortunately did violate one or more constraints but certain moment, leading to the choice of other alternative, a compromise.

Constraints are a useful tool to eliminate quickly and effectively the majority of unfit alternatives from our set of prospective solutions. They are part of sphere of the decision-maker preferences and can be added in order to a diminished amount of information and thus simplify finding of the decision problem solution. Constraints can be added or modified at any moment of the decision making modeling and process.

The other big advantage of constraints is their ability to highlight the solutions/alternatives which are for some reasons unfit or even dangerous for the sexes of the enterprise which depends on this particular decision. They allow quickly and effectively to turn focus of the decision-maker and experts to some important problems with particular solutions, which in turn can or should be mitigated or corrected,  in case the solution is preferred or acceptable for other reasons.

Constraints can be combined into groups of conditions to each of them a specific value of the weight can be assigned. The value of weight signifies the importance of the constraint amongst the other constraints in the pool. After waiting the results from revelation of the constraints in the pool the decision-maker can get additional insight of how much each alternative validates the limitations of the model. This information may suggest also that some constraints might be loosened or tightened in order to include or exclude certain alternatives from consideration. This method of revelation of constraints might also be useful to highlight that some of the methods are more or less distinguished from other alternatives that validate constraints of the model. It is an additional information about differentiation of their quality.

As it was mentioned before the constraint is the logical expression which defines the boundaries of the problem space for given decision model. This expression should be defined with use of standard logical functions like AND, OR,  NOT and the comparison operator. For more advanced users that are additional functions available like XOR, NOR, NAND etc. conditional expression can also be used to define constraint, however it is important to remember that the result of the conditional expression should always be rendered to the boolean type. the constraint expressions, especially in comparison operators, my to further tool and a input or parameter defined within the decision model by using its variable name (identifier).

Below, on the following figures has been presented an example of a condition that checks the value of the cost input. For each evaluated alternatives it checks simply in the value of the input is zero or not. If the value on the input is equal to zero the constraint is satisfied and it does not exclude this alternative from our consideration. In the reverse case case when the condition is not satisfied, which means that the constraint has been validated, the alternative is considered as unfit for further evaluation.

Simple constraint from an example the seen interview of constraints. Note that because there is only one constraint defined for this problem the weight is set to 1.

The properties pane of the simple constrained from our example. The viable "c" in our expression is the value on the input "Cost" from our decision model.

In summary, the constraints are very important tool in the hands of the decision-maker while making evaluations on large amounts of data, especially while considering numerous alternatives. By using them in proper and thoughtful manner they can lead to fast narrowing down solutions to a very small set which can be then easier and more deeply evaluated using the main evaluation method defined by the decision model.

Written by Stan Pak

March 31st, 2011 at 6:15 pm

Posted in General

Leave a Reply

You must be logged in to post a comment.