A column generation based heuristic for the generalized vehicle routing problem with time windows

The generalized vehicle routing problem with time windows (GVRPTW) is defined on a directed graph G=(V,A) where the vertex set V partitioned into clusters. One cluster contains only depot, located homogeneous fleet of vehicles, each limited capacity. other clusters represent customers. A demand associated cluster. Inside cluster, vertices possible locations customer. window vertex, during which visit must take place if visited. objective to find routes such that total traveling cost minimized, exactly one per visited, and all capacity constraints are respected. This paper presents covering formulation for GVRPTW used provide column generation based heuristic solve it. proposed solving method combines several components including construction heuristic, route optimization procedure, local search operators negative reduced routes. Experimental results benchmark instances show algorithm efficient high-quality solutions up 120 obtained within short computation times.