Under the assumption of increasing user link cost functions, the objective function (12) in
the S-O problem is convex, and the feasible set consisting of the linear constraints (7) – (9)
is also convex. Therefore, the optimality conditions, that is, the Kuhn-Tucker conditions
are: for each O/D pair ω ∈ Ω, and each path p ∈ Pω, the flow pattern x (and corresponding
link flow pattern f), satisfying (7)–(9) must satisfy: