Skip to main content

Hitachi

Corporate InformationResearch & Development

January 10, 2018

Report from Presenter


Photo 1 Meeting hall



Fig. 1 Location routing problem
Enlarge

APIEMS2017 - The 18th Asia Pacific Industrial Engineering and Management Systems – was held in Yogyakarta, Indonesia, from 3rd to 6th in December, 2017 (Photo 1). APIEMS Society encourages sharing information in the field of Industrial Engineering (IE) and Operations Management (OM) in the Asia-Pacific region. APIEMS2017 is the annual conference of the APIEMS. In the conference, there were several sessions such as logistics, maintenance and information system. Attendees from 18 countries held over 200 presentations.

We made a presentation titled "Location Routing Problem with Transportation Mode Options" which co-authored with Sophia University. The purpose of our research is improvement of the logistics efficiency. Expansion of internet shopping entails increase of delivery amount. In order to deliver efficiently, locations of distribution center and delivery routes should be decided adequately. This problem is called "the Location Routing Problem (LRP)." Problem definition of classical LRP is that a logistics provider delivers to all customer with own fleet. If there is a customer far from distribution center, delivering to the customer is inefficient (Fig. 1 left). In this case, it could be to improve delivery efficiency by entrusting the delivery to other provider (Fig. 1 right). In our study, entrusting the delivery is called as courier delivery. We expanded the classical LRP to select suitable transportation modes.


Fig. 2 Result of numerical experiments
Enlarge

In our model, the efficiency of delivery by a vehicle is compared with the efficiency of courier delivery. If the efficiency of courier delivery is higher than the efficiency of delivery by a vehicle, these customers are delivered with courier. It has become possible to select suitable transportation modes with modeling this tradeoff as a mixed integer problem. Numerical experiments using small data with less than 20 customers show that the number of customers delivered with vehicles increases as the unit price for courier delivery increases (Fig. 2). This model is a complicated combination optimization problem, and it takes more than one day to solve real scale data (about 100 customers) by a mixed integer programming solver. In the future, we will work on developing algorithms to solve the proposed model at high speed.

(By HOSODA Junko)

  • Page top