An algorithm for computing maximum solution bases (Q2641220)

The paper is concerned with families of problems defined on a common solution space, each problem having its own subset of feasible solutions. Each problem is associated with a cost and the value of a problem is the minimum cost of a feasible solution. A key concept in this approach is the maximum solution basis, which, roughly speaking is a subset of problems with the following property: finding the values of the problems in the maximum solution basis allows to compute easily the value of any of the problems. An algorithm for generating a maximum solution basis is described. The possibilities of applying this algorithm to the multiterminal cut problem are discussed.