APLIKASI SISTEM INFORMASI GEOGRAFIS, ALGORITMA DECISION TREE C4.5, FLOYD-WARSHALL, DAN 2-OPT HEURISTIC UNTUK STRATEGI PEMASARAN BARANG SECARA KELILING; APLICATION GEOGRAPHIC INFORMATION SYSTEM, DECISION TREE C4.5, FLOYD-WARSHALL, AND 2-OPT HEURISTIC ALGORITHM FOR MOBILE MARKETING STRATEGY

Rosyadi, R Ibnu, Edi Winarko

2015 | Disertasi | FMIPA

In order to improve business efficiency, a vendor company needs to precisely estimate type and quantity of items carried in each freight car, as well as determining the fastest route to the customer stores. Thus, waste of space within each car for transporting goods that are not bought and wasted fuel due to the selection of inefficient route can be avoided. This study aims to develop a system that predict the type and quantity of items to be purchased by customers using data mining method with C4.5 algorithm; to determine the fastest route to the customer stores using traveling saleman problem method with Floyd-Warshall and 2-opt Heuristic algorithm; and to present the results in a geographic information system. C4.5 decision tree algorithm using population data parameters is used to estimate the items to be carried in each freight car. The fastest sales route is obtained by determining the sequence of customer stores to be visited using 2-opt Heuristic algorithms and the fastest route between two stores using Floyd-Warshall algorithm. Impedance is used to determine the travel time routes are. The result of this study is an information system that can determine the type and quantity of items to be carried by a freight car and the fastest sales route. Both are presented in a geographic information system. By using data mining method with C4.5 algorithm, type and quantity of items to be purchased by customer and carried in each freight car can be predicted with up to 79,3% accuracy. On the other hand, route determination using 2-opt Heuristic algorithm and Floyd-Warshall algorithm tested with simulation program, produced the fastest route.

Kata Kunci : traveling salesman problem; Geographic Information Systems; C4.5 decision tree; 2-opt heuristic; floyd - warshal