KARAKTERISASI PERGERAKAN ASAL TUJUAN BARANG DI INDONESIA DENGAN PENDEKATAN JARINGAN

ARIES SETIAWAN, Bertha Maya Sopha, S.T., M.Sc., Ph.D.

2018 | Skripsi | S1 TEKNIK INDUSTRI

Balitbanghub telah melakukan survei Asal Tujuan Transportasi Nasional (ATTN), dimana survei ini mencatat pergerakan barang yang terjadi di Indonesia. Data hasil survei tersebut dapat digunakan sebagai pedoman untuk merumuskan kebijakan transportasi yang diharapkan mampu meningkatkan kinerja sistem logistik di Indonesia yang masih belum optimal. Namun, perumusan kebijakan transportasi memerlukan berbagai pertimbangan, salah satunya adalah berdasarkan pola pergerakan barang. Data hasil survei pergerakan barang masih perlu diolah untuk mendapatkan gambaran pola pergerakan barang. Penelitian ini bertujuan untuk mengkarakterisasi pergerakan barang di Indonesia berdasarkan data ATTN dengan menggunakan pendekatan jaringan. Sistem berupa jaringan yang dibentuk oleh provinsi-provinsi di Indonesia yang berperan sebagai node, sedangkan volume pergerakan barang berperan sebagai edge. Terdapat dua jenis jaringan yang dianalisis, yaitu jaringan aggregate dan jaringan disaggregate yang merupakan jaringan dari 33 kelompok barang. Analisis karakter diawali dengan perhitungan parameter jaringan dari masing-masing provinsi. Kemudian dilakukan analisis yang meliputi connectivity, assortativity, clustering, dan centrality. Selain untuk menganalisis karakter, parameter jaringan juga digunakan untuk melakukan prediksi pergerakan barang dengan metode random forest. Hasil analisis karakter menunjukkan bahwa jaringan aggregate merupakan jaringan scale-free yang bersifat diassortative dan memiliki struktur core-periphery. Jaringan aggregate tersusun dari jaringan-jaringan disaggregate yang heterogen dan saling berkorelasi positif. Walaupun heterogen, jaringan disaggregate cenderung memiliki kesamaan karakter dengan jaringan aggregate. Hasil prediksi menunjukkan bahwa similarity clustering coefficient total menjadi variabel yang paling berpengaruh pada prediksi pergerakan barang aggregate dan similarity node strength total menjadi variabel yang paling berpengaruh pada prediksi pergerakan barang disaggregate.

Balitbanghub has conducted an Asal Tujuan Transportasi Nasional (ATTN) survey, where the survey records the movement of goods occurring in Indonesia. The survey data can be used as a guide to formulate transportation policy which is expected to improve the performance of logistics system in Indonesia which is still not optimal. However, the formulation of transportation policy requires a variety of considerations, one of which is based on the pattern of movement of goods. Data from the survey of goods movement still needs to be processed to get a picture of the movement of goods. This study aims to characterize the movement of goods in Indonesia based on ATTN data using a network approach. A network system formed by provinces in Indonesia that acts as a node, while the volume of movement of goods acts as an edge. There are two types of networks analyzed, namely aggregate network and disaggregate network which is a network of 33 groups of goods. Character analysis begins with the calculation of network parameters of each province. Then performed an analysis that includes connectivity, assortativity, clustering, and centrality. In addition to analyzing characters, network parameters are also used to predict the movement of goods with random forest methods. Character analysis results show that the aggregate network is a scale-free network that is diassortative and has a core-periphery structure. Aggregate networks are composed of heterogeneous and positively correlated disaggregate networks. Although heterogeneous, disaggregate networks tend to have a common character with aggregate networks. The prediction result shows that the similarity of total clustering coefficient becomes the most influential variable on the prediction of movement of aggregate goods and the similarity node strength total become the most influential variable on the prediction of the movement of disaggregate goods.

Kata Kunci : ATTN, logistik, pendekatan jaringan, random forest