List view
- No due date•1/11 issues closed
Implement generalised Lagrange multipliers method for constrained fitting. Because the models provided to symfit are typically differentiable, the extrema of the model can be determined analytically. In cases where this is possible, this has great benefits over numerical fitting as the system can find all extrema of the model, and allows the user more control over which minimum they want. If the user provided bounds for his parameters these will be used to select the solution in this range. In cases where no analytical solution is possible, the system could default to numerical minimisation instead. The constraint syntax will be the same as for Minimize: ```python x, y = variables('x, y') a = Parameter() model = x**2 + a * y**2 constraints = [ x >= 3, x == y + 2, ] fit = Fit(model, x=xdata, y=ydata, data=zdata, constraints=constraints) ``` (I have no idea if this example makes sense, it's just to demonstrate the syntax)
No due date•2/3 issues closed