F(x) >=0

x >=0

F_i(x) x_i>=0, i=1,...,n .

If **F** depends linearly on **x**, then we have a linear, otherwise a nonlinear complementarity problem. Problems of this type occur often e.g. in mechanics, finance and games. Linear complementarity problems typically are solved by so called principal pivoting algorithms and nonlinear ones by a nonsmooth nonlinear equations approach using appropriate variants of the damped Newton's method e.g.

H(x)=0

with H_{i}(x)=SQRT( F_{i}^{2} (x)+x_{i}^{2})-F_{i} (x)-x_{i}

Systems of equations, various pieces of software, mostly in Matlab, documentation, testproblems, net-submission and other info see:

Michael Ferris' webpage |

AMPL interface to PATH |