IOSO NM- Universal software for real-life
algorithms can be purchased with
subsequent technical support and consulting help.
Could one handle 10 or 20
criteria task? The answer is "yes" - using IOSO NM Software!
Most of the real-life engineering optimization problems require simultaneous
optimization of more than one objective function.
In these cases, it is often impossible that the same values of design
parameters will lead to the optimal values for all goals. Hence, to
ensure a satisfactory design some trade-off between the objectives is
For this purpose we use the multiobjective (multigoal) approach to
optimize the overall efficiency.
The method testing
Unfortunately, currently there are no common tests for multiobjective
optimization methods. To estimate the efficiency of our method we solved
many optimization problems where we used well-known optimization test
functions as individual objectives in the Multiobjective formulation.
We made sure that the extremums of the single objectives did not coincide
with each other. The following parameters were used as a measure of efficiency
of our method:
- Precision of finding the extremes of the individual objectives;
- Average precision according to Pareto set;
- How uniform the solutions were distributed in the space of objectives.
The results obtained showed the high efficiency of our method for solving
various types of problems.