Ayarlanabilir bölme boyutları ile çok bölmeli envanter rotalama problemi

In this study, we focus on a problem where the supplier manages the customers' inventories and coordinates the distribution of multiple products to customers. Hereby the supplier can minimize the transportation costs by determining distribution routes, frequency of visits and distribution amounts simultaneously in a way that customer demands are satisfied on time. Customers place orders for each day and each product without following any pattern and have separate storage capacities for each product. Multi-compartment vehicles are used to enable different products to be distributed on a single route. We assume to have a fleet of vehicles with a certain number of unit compartments and the capacity dedicated to each product on a route can be adjusted discretely with the use of separators. This compartment structure provides the supplier with flexibility in making distribution plans, while on the other hand makes the problem more challenging since the capacity dedicated to each product on each route emerges as an additional decision to be made. To solve the proposed multi-compartment inventory routing problem we develop a novel Adaptive Large Neighborhood Search based matheuristic. We compare our performance with a benchmark algorithm we adapt from the literature by using an extensive set of instances. We observe that our solution approach outperforms the benchmark algorithm by 24.57% on average.