Multi-item single-source ordering with detailed consideration of transportation capacities

We consider multi-item single-source ordering with detailed consideration of transportation capacities. Such problems are characteristic for companies which operate direct links as part of their supply chain to transport loads with heterogeneous physical dimensions and fluctuating demands. Given knowledge on transportation demands, companies can eliminate future transports by shifting the load to fill the inflexible capacity of prior transports. While reducing transportation costs, doing so will ceteris paribus imply inventory. The problem is to coordinate orders across multiple items such that transport costs are minimized at minimal increase in inventory. The approach is distinct from prior works in that it considers detailed loading restrictions. We therefore interpret the problem as a multi-period version of the container loading problem. A wall building approach is used and incorporated into a heuristic rolling horizon procedure. We test the proposed procedure on some random problems which resemble a real inbound case from the automotive industry. As compared to period-by-period planning and two benchmarks with aggregated capacity models from the literature and practice, cost savings are possible under a wide range of operating conditions and mostly independent of the shipping volume. The largest potential exists for mid- to long-distance transports. There is a relevant potential to improve short-distance transports as well, however, only if inventory cost rates are moderate.