Searching Algorithm for an Optimal Location of Warehouses in a Distribution Network for Predicted Order Variability

Abstract

In reality, from the point of view of a logistics company, we have little control over the location of customers and producers. As a logistics company, we can, however, situate warehouses [2, 5,6,7]. The construction and location of warehouses is aimed at achieving organizational independence from production activities, to reduce transport costs, and better adapt to the temporary and overarching supply policy of recipients [1, 9, 11]. So there is an organizational buffer between the producer and the recipient. As often is the case in real-life situations, the producer or recipient has logistics departments creating so-called stocks of raw materials and distribution inventory [3, 4], but nowadays, this is not the norm. Getting back to the topic of warehouses, it is extremely important to have an optimum location. Here, we encounter a number of restrictions that take into account both infrastructural and personnel matters. Nevertheless, we can still search for an optimum location while taking restrictions into account. In reality, when looking for an optimum location one must also take into account the number of deliveries, the size, direction and delivery rate. The main purpose is to assess the usefulness and structure of network connections using the rough set theory, more specifically, several of its parameters, such as strength, degree of certainty and coverage ratio. Parameters used in works [8, 10, 12] were also used to assess risk, reliability, security, etc. Based on the estimators of selected parameters, one can create a standard set of rules and use them in the inference mechanism to categorize network connections and determine their usefulness. The proposed concept is multi-tiered, which corresponds to the individual sections of the publication. The final goal consists in using the rough set theory to make predictions (and thus a structural proposal) for a treated (functionally better than the current one) configuration of the distribution network with an optimum (or near optimum) warehouse location.