Applies to:
@RISK 5.x and Up, Industrial Edition 
Evolver 4.x and Up
RISKOptimizer 1.x
RISKOptimizer Developer's Kit (RODK) 4.1
Evolver Developer's Kit (EDK) 4.1

How to Speed Up Optimization in RISKOptimizer or Evolver

Issue:
Your optimization seems to take a long time to execute. There are several strategies to improve performance and reduce runtime.

Checklist for Faster Optimization:

1. Upgrade to the Latest Version

   • The performance of the OptQuest engine in @RISK Industrial has improved over time.

2. Choose the Appropriate Solving Method

   • Limit adjustable cells to a small range to increase valid trials.

   • Example: Assigning numbers 1–20 to 20 cells—use Order rather than Recipe to avoid generating duplicates that must be eliminated by constraints.

3. Use Hard Constraints Appropriately

   • Evolver and RISKOptimizer handle hard constraints efficiently.

   • The OptQuest engine and Genetic Algorithm intelligently search for solutions that meet constraints.

4. Make Constraints Linear

   • Linear constraints reduce invalid trials and improve optimization speed.

   • Avoid using MAX or MIN functions; constrain the cell range directly.

   • OptQuest performs best when = constraints are linear.

5. Adjustable Cells: Use Discrete or Integer Values

   • Discrete adjustable cells allow OptQuest to enumerate possibilities and generate only valid trials.

6. Start with a Feasible Solution

   • Especially important for the Genetic Algorithm. Begin with all constraints met to reduce search time.

7. Optimize Continuous Target Values

   • Target cells should respond gradually to changes in adjustable cells.

   • Avoid binary targets (1/0) that provide little guidance; use constraints to rule out unacceptable solutions.

8. Constrain Natural Continuous Values

   • Example: If C5 must be ≤120, set the constraint directly rather than encoding it in other logic.

9. Enable Multi-Threaded Calculations

   • Excel 2010–2016: File → Options → Advanced → Formulas → Enable multi-threaded calculations

10. Use Optimization Stopping Conditions

    • Stop when adequate convergence is reached; the last portion may offer little improvement but consume much time.

    • In RISKOptimizer, also set simulation stopping conditions on the Convergence tab of @RISK Simulation Settings (6.x/7.x) or on the RISKOptimizer Options screen (1.x/5.x).

11. Speed Up Simulation Portion

    • Follow @RISK best practices for faster simulations.

    • If no @RISK distribution functions are used, set iterations to 1 or consider using Evolver.

12. Leverage Multiple CPUs (RISKOptimizer 7.5+)

    • Ensure multi-CPU settings are Automatic or Enabled.

    • For better performance, use a machine with more cores and ample RAM.

Summary:
Optimization speed can be improved by using the latest software version, selecting the right solving method, applying hard and linear constraints, starting with feasible solutions, optimizing continuous values, enabling multi-threading, and leveraging multiple CPUs.

 

Related Articles:

@RISK: Speed Up Performance

 RISKOptimizer: Running RISKOptimizer Deterministically