Logic cuts for multilevel generalized assignment problems

摘要: Abstract In the multilevel generalized assignment problem (MGAP) agents can perform tasks at more than one efficiency level. Important manufacturing problems, such as lot sizing, be easily formulated MGAPs; however, large number of variables in related 0–1 integer program makes it hard to find optimal solutions these even when using powerful commercial optimization packages. The MGAP includes a set knapsack constraints, per agent, that useful for generating simple logical constraints or logic cuts. We exploit fact cuts generated linear time and added model before solving with classical branch bound methodology. generate all contiguous 1-cuts every MGAP’s problems report effect adding our experimental results.