Monday, October 1, 2012

Analysis and Improvement of the Product Delivery System of a Beer Producer in Ankara


Customer Demand of BEER!
 

Geographical Information Systems (GIS) are often seen to many as having limited use as only having the capability to map the geographical landscape and its environmental characteristics. However, it is interesting to see that GIS can also play a very beneficial role in the solving of large-scale logistic problems as well. To solve these large-scale logistic problems one must first look into a combined analysis of spatial, non-spatial, and qualitative data as factors involved with these problems. In this article the authors, S. Pamuk, M. Köksalan and R. Güllü, are hired by a major beer producer in Turkey to analyze and improve his current beer delivery system. To do this, the authors outline their analysis of these factors involved with large-scale logistic problems and provide for the practicality and utility of implementing a GIS project in order to produce improvements that can be implemented to the current problematic beer delivery system. 

The authors use Ankara, Turkey as the pilot city for this study as it is the capital and second most crowded city of Turkey, in which their client has some 4000 customers who are supplied with canned or bottled beer from a single depot. Before the authors’ involvement, this beer delivery system accepted orders from each customer twice a week and any customer placing a nonzero order received its beer on the workday following the order. The orders were divided by delivery day pairs that were assigned by geographical location: Monday – Thursday, Tuesday – Friday, and Wednesday – Saturday. The delivery schedule was further divided by a dozen geographical districts in which the daily routing decisions were made independently. These routing decisions were based on road conditions, preferred hours of delivery for each customer, and the cases of empty bottles that were to be collected; from which the truck crews would determine what route to take and group the customers’ orders into truckloads according to that route. From this initial beer delivery system, the authors found several weaknesses:  “many customers habitually placed very small orders (1-2 cases) at each visit; varying workloads for different workdays resulted in persistent idle time/overtime costs, indicated the city’s partitioning needed to take additional criteria other than geography into account; and The existence of independent districts could occasionally cause trucks to be loaded inefficiently” (Pamuk, Köksalan & Güllü, 2004). This analysis of the current beer delivery system allowed the authors of this article to determine what was needed to improve performance and reduce cost.

The solution to the client’s large-scale logistic problem was derived from the formation of an equation that took into account all aspects of the delivery system.  The most important aspect of their equation was the separation of one-delivery-per-week and two-deliveries-per-week categories depending on average demand per week; where one-delivery-per-week customers were assigned to receive their shipment a single day of the week Monday through Saturday, and two-deliveries-per-week customers were assigned to receive their shipment on one of the day pairs Monday – Thursday, Tuesday – Friday, or Wednesday – Saturday. Other aspects of the equation included the fixed estimate of time allotted to each visit (which increased if located in city center), the estimate of time associated with unloading each unit of beer, different values for variability in the schedule of deliveries, and the distance of customer to representative cluster point. This equation was a result of the proposed improvements of which the authors recommended to be implemented to the beer delivery system to increase efficiency and reduce costs. These included:  “Determine low-demand customers to be visited less frequently than twice a week; reconstruct the three partitions in a way that would balance the average workloads of the six workdays; handle special zones such as the center zone in Ankara separately, to ensure that their workloads would also be balanced over workdays; eliminate districts and construct a single sequence of points for each zone-workday combination; and use GIS to obtain coordinate and distance data for the clustering and sequencing problems, and also to display the results to loading experts, who could ‘repair’ or change them as necessary” (Pamuk, Köksalan & Güllü, 2004). The authors used the system of Euclidean distances in their application of the equation to different customer types, which include their location within the city, scheduled deliveries, and demand types, when creating a delivery sequence. This equation allows for the client to organize his customer’s deliveries into different zone-day combinations in which different zones’ routes for the same day could be treated together during the loading phase to increase efficiency.  With the help the their university’s  educational GIS and basemaps, the authors use this equation to create and implement a new efficient delivery system sequence that can be easily understood and followed through visual displays of customer maps like figure 4 and sequence maps like figure 5 below. Figure 4 maps customers that need to be visited on Monday with the darker stars representing one-visit-per-weak customers and the lighter stars representing two-visits-per-week customers, and figure 5 maps out a route that is the most time and cost efficient for delivery to those customers.

Along with this equation created to address and improve the factors of large-scale logistic problems in the most successful way possible, our authors also find GIS to be possibly very beneficial to solving problems of this kind. They feel that GIS will not only be helpful in the problem studied in this article but also in logistics, marketing, and monitoring issues. For example, “locating new facilities, partitioning/segmentation of markets, analyzing the geographical distribution of sales versus demographic data or strategic locations, identifying low-demand areas and potential markets, real-time monitoring of vehicle movements through GPS devices, and immediate handling of distribution problems (vehicle breakdowns/accidents, etc.)” (Pamuk, Köksalan & Güllü, 2004). As for GIS improving this client’s problem, it can easily extract coordinate or zone information and distance information using any kind of metrics needed, address local traffic conditions and vehicle restrictions easily, and automatically includes the extracted distances of what you are mapping compared to having to calculate it using something like Euclidean distances as the authors did in their equation. However, GIS does have its problems and limitations especially for developing nations. A big setback in the use of GIS in developing nations is the lack of access to digitalized geographical information that may be bought at a reasonable cost, as well as the overall initial capital needed for the program itself. Even with the authors’ cheap access to GIS applications due to their university’s educational GIS program and basemaps provided through that, they still found many issues with using GIS in developing countries, in this case Turkey. These issues were “the lack of established address conventions and the unreliability of existing databases of non-spatial data” (Pamuk, Köksalan & Güllü, 2004).  The geocoding needed to appropriately use GIS applications only automatically found no more than twenty percent of the records needed in their application. Outside of the geocoding, around 300 customers of the 4400 total were problematic due to duplicate records, incomplete addresses, addresses in other cities, and so on. However, in the end, the authors of this article were eventually able to locate the majority of the customers on the digital maps with sufficient accuracy by locating the streets of the customers and manually positioning the customers that were on that same street.

By the end of this analysis of the client’s beer delivery system in Ankara, Turkey, the authors of this article were able to provide how to improve this system by 10-25 percent in distribution costs. The use of GIS in this analysis and improvement was greatly helpful to the authors and would continue to be immensely useful to the client if he were to invest the capital in the GIS program and the digitalized geographical information that is needed to run successful GIS applications. In this way, the need for an ‘information infrastructure’ is huge in order to allow for everyday successful GIS use in developing countries. It is interesting to see how GIS has been helpful in solving large-scale logistic problems in Turkey, and I believe GIS can be much more helpful to the United States in solving these same types of problems in our private business sector as we do have a solid ‘information infrastructure’  with cheap or free access GIS applications and materials. I expect to see much greater use of GIS in solving large-scale logistic problems in the United States, as well as in many other facets of our American Society as it is much more available to us for use and can be applied to so many different venues of our civilization.

Work Cited
Pamuk, S., Köksalan, M., & Güllü, R. (2004). Analysis and Improvement of the Product Delivery System of a Beer Producer in Ankara. The Journal of the Operational Research Society, 55(11), 1137-1144. Retrieved from http://www.jstor.org/stable/4101885

4 comments:

  1. Thanks for this article, it really highlights how many business use GIS in logistics. What is interesting about logistics is that distance = money. So in this case GIS is awesome at finding efficiencies in the delivery system. What is very interesting about logistics is that improving systems to reduce 10-25 percent of distribution costs, also reduces the environmental impact of the company. So in this case saving money is green, which is really cool. From a policy point a view this could be something regulation can foster. For example making it part of commercial driver's licence to perform an efficiency analysis. This way companies save money on costs, but the environmental impact is also reduced.

    ReplyDelete
  2. Awesome. Using GIS to supply more beer more efficiently. In all serious tho, it makes sense that a business would want to map the most efficient method in supplying their goods. I wonder if GIS takes traffic into account?

    ReplyDelete
  3. I think this is a perfect idea of how GIS fits into logistics. It dovetails perfectly because logistics already requires knowledge of a map and data, so GIS is just another way of organizing the data and making it more efficient.

    ReplyDelete