Optimization with solvers

At a glance

Solvers automatically find the best configuration to achieve an objective under constraints. Define what you want to optimize (minimize a cost, maximize a margin), the adjustable variables, and the constraints to respect. The solver explores the possibilities and proposes the optimal solution.

Before you begin

• Entity types with instances and KPIs must exist in your Ontologie.
• You need the nodes.write scope to create and run solvers.
• A scenario can be used as a starting point.

Steps

Configure a solver

1. Open the Simulation module from the sidebar.
2. Select the Solvers tab.
3. Click New solver.
4. Define the optimization problem:

Parameter	Description	Example
Objective	What you want to optimize.	Minimize total procurement cost
Direction	Minimize or maximize.	Minimize
Target KPI	The indicator to optimize.	total_cost
Decision variables	The properties the solver can adjust.	order_quantity, delivery_frequency
Constraints	The limits to respect.	max_budget <= 100000, delivery_time <= 5 days

5. Click Save.

Run the optimization

1. Click Run on the configured solver.
2. The solver runs iterations to converge toward the optimal solution.
3. Follow the progress:
   • Completed iterations: the number of iterations performed
   • Best value: the current value of the objective
   • Satisfied constraints: the number of constraints met

Analyze results

Once the solver is finished:

1. Review the Results section:

Section	Description
Optimal solution	The best configuration found by the solver.
Objective value	The final value of the optimized KPI.
Constraints	The status of each constraint (satisfied or violated).
Justification	The explanation for the chosen solution.
Proposed mutations	The modifications to apply to achieve this solution.

2. Compare the optimal solution with the current situation:
   • Before: current values of the variables and KPI
   • After: values proposed by the solver
   • Gain: the improvement in absolute value and percentage

Apply the solution

If the solution suits you:

1. Click Create a scenario from the proposed mutations.
2. A new scenario is created with the solver's modifications.
3. You can then apply this scenario to production.

tip

First create a scenario from the solver results, then evaluate the full impact before applying. This allows you to verify cascading effects.

Expected outcome

The solver has identified the optimal configuration for your objective, while respecting the defined constraints. The proposed mutations are ready to be transformed into a scenario for evaluation and application.

Limitations and common errors

Situation	Solution
"No solution found"	The constraints may be too restrictive. Relax one or more constraints and try again.
The solver runs too long	Reduce the number of decision variables or increase the tolerance.
"Constraint violated"	The solver could not satisfy all constraints. Review the details to identify the conflicting constraint.
The proposed solution seems incorrect	Verify the objective and constraint definitions. A reversed direction (minimize instead of maximize) is a common mistake.

Need help?

Contact us: Support and contact.