²For those who are really interested in optimizing circuits: Some parameters are very important for quick and correct convergence. The optimizer walks step by step to a local minimum of the cost function you define. Starting from an initial vector YOU provide, it converges step by step. Consider trying another start vector if the result is not the one you expected.

The optimizer will carry on walking until it reaches a vector which resulting cost is smaller than the target cost YOU provide it. You will also provide a maximum iteration count in case the target can not be achieved. Balance your time, specifications, and every other parameters. For a balance between quick and accurate convergence adjust the "factor" variable, at the beginning of minimumSteepestDescent in the file differentiate.tcl.