MAGNET Suite

OptiNetAutomated Design Optimization

OptiNet is an automated design optimization option to MagNet, ElecNet and MagNet-ThermNet coupled together. Using advanced and efficient algorithms, OptiNet can find optimal values for different design variables within the constraints specified.

OptiNet's useful features include:

  • Continuous-value and discrete-value variables and optimization
  • Evolutionary-based Stochastic search is very efficient, even for a large number of parameters
  • Built-in and customizable scripts for objective functions and constraints
  • Evaluate the impact of variations in the design parameters

OptiNet offers an integrated Automated Optimal Design environment compatible with the industrial design process by meeting the following requirements:

Robust

Independent of the number of design variables, problem type and objective

Global Search Oriented

Identifies the global minimum of an objective function

Derivative-Free

OptiNet avoids errors by using proper functions and not their derivatives

Fast

Highest possible efficiency with freedom to choose between accuracy and computing time

Features

Design Variables
  • Typically performance requirements, component & physical limitations and production-based limitations amongst other factors
  • Parameters from models in MagNet, ElecNet or ThermNet are identified as design variables by OptiNet
  • Design variables are one of four types
    • Constant
    • Continuous
    • Discrete step
    • Discrete list of values
Constraints
  • Weighted constraints including absolute type which must be satisfied
  • Model-based and Solution-based constraints
  • Common constraint types are pre-defined
  • User-defined constraints supported
Optimization Objective
  • Supports weighted multiple objectives for a design problem
  • Model-based and Solution-based objectives
  • Common types of objective functions are included
  • User-defined optimization objectives supported
Reporting Features
  • All attempted permutations are reported as improved, not improved or failed
  • Information reported includes the values of the variables, constraints & objectives
  • Create an animation of the metamorphosis from initial to optimized device
  • View all solutions computed or only solutions with improved results