Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Convex set for similarity constraint

  1. Jul 18, 2011 #1
    I am trying to ultimately find the projector onto a convex set defined in a non-explicit way, for a seismic processing application.

    The signals in question are members of some Hilbert Space H and the set membership requires that they must correlate with each other above some scalar [itex]\rho[/itex], given that the known signal [itex]\textbf{w}[/itex] is in the set. Symbolically, I want to find a projector [itex]\textit{P}[/itex] onto the convex set [itex]\textit{C}[/itex]:


    C = \left\{\mathbf{u}(t) : \left\langle \hat{\mathbf{u}}(t),\hat{\mathbf{v}}(t) \right\rangle \geq\rho_{0}, \forall \mathbf{v}(t) \in C, \quad where \quad \mathbf{w}(t) \in C \right\},


    Any intermediate help is appreciated, i.e., is there an equivalent way to formulate this set, that make finding the projector easier?
    Last edited: Jul 18, 2011
  2. jcsd
  3. Jul 18, 2011 #2


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I'm a bit confused by how C references itself in its definition.
  4. Jul 18, 2011 #3
    I understand the confusion--that is what makes the defining characteristic a little awkward. The basic idea is this: each element in the convex set [itex]\textit{C}[/itex] must correlate with every other element above [itex]\rho[/itex]. But we also know that a given (known) element [itex]\textbf{w}(t)[/itex] is contained in [itex]\textit{C}[/itex]. Is that less confusing of a statement?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook