Customers’ classification for pick-up’s demand by using the K-means clustering: A case study of urban freight transportation in Casablanca city
Autor: | Lissane Elhaq Saad, Bourrich Leila |
---|---|
Rok vydání: | 2019 |
Předmět: |
Upstream (petroleum industry)
Operations research Process (engineering) Computer science 020209 energy k-means clustering Global vision Management model 02 engineering and technology Vehicle routing problem 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Routing (electronic design automation) Cluster analysis |
Zdroj: | ISC2 |
DOI: | 10.1109/isc246665.2019.9071677 |
Popis: | the pick-up of parcels presents the upstream link of the urban freight transportation chain, it is based on pick-up routes by responding to customer requests. Each transport company has its own management model. For this link, whose main objective is to maximize pick-up requests and satisfy a very demanding clientele. In Morocco, this process knows a large number of difficulties due to several criteria linked to customers satisfaction’s factors. These difficulties make the transportation companies losing every year an important rate of their customers. In literature, this problem has not been studied before, for this reason our objective is the proposition of an efficient model that helps to know a global vision on the customer’s behavior in relationship with this kind of transportation. Then the prediction of the pick-up requests for each customer during a specific period to finally optimize the routing vehicle problem associated to our case study. In this paper, we explore the effect of the classification of customers per categories, by using the K-means clustering method. This solution will help to group our dataset composed by 256 customers from Moroccan Transportation Company. Where their kind of demand is intermittent. This classification is into N clusters with specific characterizes. Thanks to this method, we find as a result that the number of cluster K=4 is the optimal classification of our dataset and their gathering by cluster which each one had specific characteristics. |
Databáze: | OpenAIRE |
Externí odkaz: |