Mixed Integer Linear Programming (MILP) from Discrete Optimization can be applied to a vast number of problems and areas. In the meanwhile, the solvers for MILPs have progressed so considerably that Integer Programming is often superior to developing dedicated algorithms. MILP has many applications, most prominently in optimizing time tables and the traveling salesman problem. The beauty of this method is that it can employ a given list of rules or constraints limiting the search space on top of which the optimization is performed. In biochemistry, stoichiometric equations have been assembled for several thousand metabolites and reactions restricting the space of possible metabolic fluxes (at steady state) leading to Flux Balance Analysis as a MILP application1. We used MILP to optimize arrangements of cell-networks for identifying distinctively expressed pathways2 or to infer gene regulation3. Others used MILP to infer gene regulatory modules in cell-networks4 or regulatory pathways5. Within the tutorial, we will give an introduction into the principle idea of MILP, give some prominent problems to solve (such as knapsack, Sudoku) and will explain some of the above mentioned applications in bioinformatics. The participants will use their own laptops and an R-Studio and (academically free) solver (Gurobi) installation.
