Memetic algorithms for solving a bi-objective Transportation Location Routing Problem

In this work we propose two memetic algorithms based on a Nondominated Sorting Genetic Algorithm II (NSGA-II) scheme for solving a bi-objective Transportation Location Routing Problem (TLRP). The TLRP we are dealing with, can be seen as an extension of the two stage Location Routing Problem, with a transportation problem with truck capacity as first stage. Two objectives are considered: minimizing total distribution cost and balancing workload for drivers in the routing stage. The memetic algorithms we propose are attempts to reduce the typical intrinsic randomness of genetics methods. In the first memetic, instead of the classic mutation operator, a simple local search is implemented; in the second algorithm, a Tabu Search procedure with short term memory is used as mutation operator and the stop criterion is based on the quality of the solutions instead of a fixed number of generations. Computational Results shown that second algorithm is superior in four of the five performance measures applied.