Henrik Andersson

ReLU Networks as Surrogate Models in Mixed-Integer Linear Programs

  • Henrik Andersson
  • Bjarne Grimstad
    • Categories (Mixed) Integer Linear Programming

    We consider the embedding of piecewise-linear deep neural networks (ReLU networks) as surrogate models in mixed-integer linear programming (MILP) problems. A MILP formulation of ReLU networks has recently been applied by many authors to probe for various model properties subject to input bounds. The formulation is obtained by programming each ReLU operator with a binary … Read more

    Global Optimisation of Multi-Plant Manganese Alloy Production

  • Henrik Andersson
  • Stein Wasbø
  • Martin Digernes
  • Brage Rugstad Knudsen
  • Lars Rudi
  • Magnus Stålhane
    • Categories Global Optimization Applications, Production and Logistics, Supply Chain Management

    This paper studies the problem of multi-plant manganese alloy production. The problem consists of finding the optimal furnace feed of ores, fluxes, coke, and slag that yields output products which meet customer specifications, and to optimally decide the volume, composition, and allocation of the slag. To solve the problem, a nonlinear pooling problem formulation is … Read more