User manual PALISADE RISKOPTIMIZER 5.5

[. . . ] Guide to Using RISKOptimizer Simulation Optimization for Microsoft Excel Version 5. 5 May, 2009 Palisade Corporation 798 Cascadilla St. Ithaca, NY USA 14850 (607) 277-8000 (607) 277-8001 (fax) http://www. palisade. com (website) sales@palisade. com (e-mail) Copyright Notice Copyright © 2009, Palisade Corporation. Trademark Acknowledgments Microsoft, Excel and Windows are registered trademarks of Microsoft, Inc. IBM is a registered trademark of International Business Machines, Inc. Palisade, RISKOptimizer, TopRank, BestFit and RISKview are registered trademarks of Palisade Corporation. [. . . ] Each group of adjustable cells should contain unique adjustable cells because the values in the first group of adjustable cells would be ignored and overwritten by the values in the second group of adjustable cells. If you think a problem needs to be represented by more than one solving method, consider how to break up the variables into two or more groups. Chapter 5: RISKOptimizer Reference Guide 105 Constraints RISKOptimizer allows you to enter constraints, or conditions that must be met for a solution to be valid. Constraints you have entered are shown in the Constraints table in the Model Definition dialog box. Add - Adding Constraints Clicking the Add button next to the Constraints table displays the Constraint Settings dialog box where constraints are entered. Using this dialog box the type of constraint desired, along with its description, type, definition and evaluation time can be entered. 106 Model Definition Command Constraint Type Two types of constraints can be specified in RISKOptimizer: · Hard, or conditions that must be met for a solution to be valid (i. e. , a hard constraint could be C10<=A4; in this case, if a solution generates a value for C10 that is greater than the value of cell A4, the solution will be thrown out). Soft, or conditions which we would like to be met as much as possible, but which we may be willing to compromise for a big improvement in fitness or target cell result (i. e. , a soft constraint could be C10<100; however, C10 could go over 100, but when that happened the calculated value for the target cell would be decreased based on the penalty function you have entered). · Evaluation Time Hard constraints may be evaluated 1) each iteration of a simulation run for a trial solution (an "iteration" constraint), or 2) at the end of the simulation run for a trial solution (a "simulation" constraint). An iteration constraint is a constraint that is evaluated each iteration of a simulation run for a given trial solution. If an iteration results in values which violate the hard constraint, the simulation is stopped (and the trial solution rejected) and the next trial solution and its associated simulation begins. A simulation constraint is specified in terms of a simulation statistic for a spreadsheet cell; for example the Mean of A11>1000. A simulation constraint, as opposed to an iteration constraint, will never cause a simulation to be stopped prior to completion. · Chapter 5: RISKOptimizer Reference Guide 107 Simulation Constraints A simulation constraint is specified in terms of a simulation statistic for a spreadsheet cell; for example the Mean of A11>1000. The statistic to use in the constraint is selected from the available dropdown list: When a simulation constraint is used, a distribution of possible values for the Range to Constrain is generated during each trial solution's simulation. At the end of each simulation, the constraint is checked to see if it has been met. If the simulation constraint is a hard constraint and the constraint is not met, the trial solution is discarded. If the constraint is a soft constraint and the constraint is not met, the target cell's statistic that is being minimized or maximized is penalized according to the entered penalty function (see the next section Soft Constraints). 108 Model Definition Command Simple and Formula Constraints Two formats ­ Simple and Formula -- can be used for entering constraints. The type of information you can enter for a constraint depends on the format you select. · Simple Format - The Simple format allows constraints to be entered using simple <, <=, >, >= or = relations where a cell is compared with an entered number. A typical Simple constraint would be: 0<Value of A1<10 where A1 is entered in the Cell Range box, 0 is entered in the Min box and 10 is entered in the Max box. With a simple range of values format constraint, you can enter just a Min value, just a Max or both. The entered Min and Max values must be numeric in the simple range of values constraint format. · Formula Format - The Formula format allows you to enter any valid Excel formula as a constraint, such as A19<(1. 2*E7)+E8. RISKOptimizer will check whether the entered formula evaluates to TRUE or FALSE to see if the constraint has been met Soft Constraints Soft Constraints are conditions which we would like to be met as much as possible, but which we may be willing to compromise for a big improvement in fitness or target cell result. When a soft constraint is not met it causes a change in the target cell result away from its optimal value. The amount of change caused by an unmet soft constraint is calculated using a penalty function that is entered when you specify the soft constraint. Chapter 5: RISKOptimizer Reference Guide 109 More information about penalty functions is as follows: · Entering a Penalty Function. [. . . ] Such random numbers are the basis for other routines that convert them into samples drawn from specific distribution types. See random sample, seed A random sample is a value that has been chosen from a probability distribution describing a random variable. Such a sample is drawn randomly according to a sampling "algorithm". The frequency distribution constructed from a large number of random samples drawn by such an algorithm will closely approximate the probability distribution for which the algorithm was designed. Population Probability Probability Distribution Random Number Generator Random Sample 214 Ranges In RISKOptimizer: The user sets the range, or the highest and lowest value that RISKOptimizer is allowed to try when adjusting a certain variable. [. . . ]