Distribution planning for products with varying life cycles

Distribution planning becomes increasingly complex when warehouses handle products with varying life cycles; long (basic) and short (fashion). The varying demand patterns and life-cycles associated with each product class require different planning decisions possibly with different objectives; cost-efficiency for basic products and time-responsiveness for fashion products. Although it may appear that such products may be handled separately from inventory and transportation perspectives, they certainly are inseparable at the warehouse as they often get handled simultaneously by the same warehouse resources (i.e., workers and technology). To address this challenge, we introduce a novel problem to the literature, the two-product class, multi-product quantities, multi-echelon, multi-period, distribution planning problem. The uniqueness of this problem is in jointly deriving optimal distribution plans for both product classes, combined with optimal warehouse design decisions (e.g., technology and workforce), in order to minimize the total cost. We model this problem as a mixed-integer bilinear optimization problem, which turns out to be computationally challenging when dealing with industry-size instances (e.g., 200 stores, 1000 products, and 28 time-periods). We, therefore, propose a heuristic based on the iterative local search framework and conduct a comprehensive experimental study. Results suggest that with an increase in the proportion of fashion products, there is a substantial increase in warehouse workload variation, and subsequently total cost. An increase in fashion window tends to mitigate the variation and so do situations where the warehouse is located in regions with higher worker costs.