Replacing swarm members that fail

I recently asked this question: http://uk.mathworks.com/matlabcentral/answers/264835-genetic-algorithm-fitness-function-for-failing-parameters

In it I am attempting to optimise a parameter set for a strongly nonlinear model using GA. Some of the parameter sets create models which fail, which I then want to replace with working sets. The solution was to use the output function to generate random new sets within the parameter bounds until one provided a working solution, and to insert this into the generation.

I am now attempting to do the same with Particle Swarm Optimisation, but with PSO, the output function does not allow changing of the swarm. Is there another way to edit the swarm members?