A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Traveling Salesman Problem on Smart Meter Infrastructure
Delivery route optimization is one of the possible applications on smart meter infrastructure. On the premise of security and user's agreement on data usage, electricity data from smart meter enables occupancy prediction for absent delivery avoidance. However, it is necessary to solve two technical issues to realize this. The occupancy prediction needs to be applicable without collecting the occupancy labels. The other is to calculate the shortest route while avoiding the absent destination using a given occupancy probability. For occupancy prediction, this study evaluated the generalization performance of proposed methods by collecting electricity consumption data and occupancy labels for 23 households in addition to the existing dataset. Then, we conducted actual parcel deliveries with delivery operators on the proposed system to 150 households. While it succeeded in reducing 50% of absent delivery in a real-world environment, total delivery distance increased. Thus, we constructed a new problem formulation, “Traveling Salesman Problem with Occupancy Prediction (TSP with OCP),” to solve route optimization problems considering occupancy probability and absent delivery cost. As a result of the simulation, TSP with OCP showed a 16% improvement in total distance than TSP without OCP.
Traveling Salesman Problem on Smart Meter Infrastructure
Delivery route optimization is one of the possible applications on smart meter infrastructure. On the premise of security and user's agreement on data usage, electricity data from smart meter enables occupancy prediction for absent delivery avoidance. However, it is necessary to solve two technical issues to realize this. The occupancy prediction needs to be applicable without collecting the occupancy labels. The other is to calculate the shortest route while avoiding the absent destination using a given occupancy probability. For occupancy prediction, this study evaluated the generalization performance of proposed methods by collecting electricity consumption data and occupancy labels for 23 households in addition to the existing dataset. Then, we conducted actual parcel deliveries with delivery operators on the proposed system to 150 households. While it succeeded in reducing 50% of absent delivery in a real-world environment, total delivery distance increased. Thus, we constructed a new problem formulation, “Traveling Salesman Problem with Occupancy Prediction (TSP with OCP),” to solve route optimization problems considering occupancy probability and absent delivery cost. As a result of the simulation, TSP with OCP showed a 16% improvement in total distance than TSP without OCP.
Traveling Salesman Problem on Smart Meter Infrastructure
Ohsugi, Shimpei (author) / Negishi, So (author) / Okada, Kensaku (author) / Yoshii, Hayato (author) / Tanaka, Kenji (author) / Koshizuka, Noboru (author)
2021-09-07
6561724 byte
Conference paper
Electronic Resource
English
Perancangan Aplikasi Penjadwalan Traveling Salesman Problem dengan Algoritma Genetika
| DOAJ | 2012
Research on Traveling Salesman Problem of Short-Haul Freight Transportation
| British Library Conference Proceedings | 2013
New Formulation for Traveling Salesman Problem with Separation Requirement and Cost
| British Library Online Contents | 2011