Talk:Active-set method

In describing the general structure of an active set algorithm, I am not clear about several points.

The first point is "remove a subset of the constraints with negative Lagrange multipliers". I guess it a subset from the established active set since the Lagrange multipliers exist only for elements in the established active set.

I cannot understand the last point either: "search for infeasible constraints". Is it a separate step or a step logically connected with its previous step?

If it is a separate step, that means, I switch my attention away from active set, but check whether the left-over constraints which are not included in the active set can be satisfied with the newly calculated x, then there should be more wording, I guess, and there should be comment on what to do with those constraints fulfilled by these x.

If it a logically continued step, then I would ask what actually are supposed to be done here. Since the previous step withdraws several active members out of the set, are we actually checking to see whether any constraints outside the active set can be included into the active set, otherwise, members in the active set would keep decreasing.

Please clarify, thanks,

True bsmile (talk) 22:56, 19 October 2009 (UTC)true_bsmile