Optimization of warehouse order-picking routes using vehicle routing model and genetic algorithm

Order-picking is one of the most time-intensive and costly processes in any warehouse. Accelerating the order-picking process not only increases revenue by delivering more product volume per unit time but also reduces labor cost, thereby increasing prot on two fronts. With the overall volume and variety of orders increasing, it is now more important than ever for warehouses to be able to pick orders quickly. Since traveling accounts for a significant portion of an order-picker's time, optimizing the routes they take can have tremendous positive effects on overall warehouse costs. In this research, a warehouse with multiple zones is studied. The storage policy is a low-level forward-reserve allocation with dedicated pick slot storage. The picking policy is manual picker-to-part and pick by order. An actual sample of orders was collected and re-batched and the picking process is modeled as a capacitated vehicle routing problem. The objective is to minimize the total travel distance among all order-pickers. The model is solved using a genetic algorithm developed in MATLAB. The algorithm utilizes single point crossover and simple random mutation operators. The main parameters of the algorithm are the population size, probability of crossover, and mutation rate. The algorithm is run using many parameter combinations for a sample problem as well as for each zone. The algorithm was able to obtain the optimal solution multiple times for the sample problem. For the smallest sub-problem, the solution was 6% away from optimal. The grand total travel distance for all routes in all zones was 10,468.2 feet. The algorithm quickly finds high-quality, practical solutions that could potentially be implemented into actual warehouses to reduce order-picking time and overall warehouse costs.