Isti Surjandari

Petrol Delivery Assignment with Multi-Product, Multi-Depot, Split Deliveries and Time Windows

The purpose of this study is to solve a planning problem faced by Pertamina, a state-owned Oil and Gas company in Indonesia, dealing with the delivery of petrol products. Based on Petrol Station Replenishment Problem (PSRP), a fleet of tank-trucks with different capacities has been assigned to deliver two petrol products from two depots with split deliveries and time windows to a set of petrol stations in the working area. The Tabu Search (TS) algorithm has been used to solve this problem. The problem consists of jointly determining the fleet assignment and of designing delivery routes to petrol stations to satisfy the orders using the available resources with the minimum total travel cost for delivery.

Identifying Consumer Buying Behavior Differences through Market Basket Analysis in Multiple Outlet Types

Retail business in Indonesia is growing rapidly and it causes a sales competition conducted by retailers. Therefore, a strategy is needed to expand the consumer market. In order to expand the current market, companies need to pay attention to customer satisfaction that will affect the sustainability of product purchasing activities. Market basket analysis is done to extract consumer buying behavior by identifying the associations of various products that consumers put on the shopping basket. This study was conducted to observe whether outlet type affects consumer buying behavior. There are two types of outlets based on their placement locations, which are outlets located within shopping centers and stand-alone outlets. The data used in this study was taken from customer purchase transactions database. Since the purchase transactions made by consumers were large, then data was processed using data mining techniques. Apriori algorithm was applied for finding frequent item-set. The minimum support used in data processing was 12%, while the minimum confidence used was 70%. Based on these results, it can be concluded that the pattern of consumer purchases in outlets within the shopping center was making a purchase with a few number of product category combinations, while in stand-alone outlets were engaged in transactions with the number of combinations of more product categories. The difference can be observed from the association pattern generated from each type of outlet. By knowing the pattern of consumer purchases at each outlet viewed from the product association, the company can determine the right strategy to increase the market for each type of outlet.

Forecasting Methods for Determining the Level of Safety Stock in Electronic Industry

The competition in industries force many companies to reduce production costs in order to gain more profit. This decision requires a good planning for production schedule. In production planning, forecasting theory to estimate future demand is an important aspect. The problem in this article is to choose the best forecasting method based on the characteristics of historical data in seasonal demand patterns. Results forecast error is then used to calculate the safety stock as a form of anticipatory strategies shortage of material. The results indicate that the best forecasting method is a combination method of Holt-Winters Exponential Smoothing and Naive, it can decrease Mean Absolute Percentage Error (MAPE) up to 63%.

Oil fuel delivery optimization for multi product and multi depot: the case of petrol station replenishment problem (PSRP)

With the Oil and Gas Law No. 22 of 2001, national and foreign private enterprises can invest in all sectors of Oil and Gas in Indonesia. In anticipation of this free competition, Pertamina, as a state-owned enterprises, which previously had monopolized the oil and gas business activities in Indonesia, should be able to improve services as well as the efficiency in order to compete in the free market, especially in terms of cost efficiency of fuel distribution to gas station (SPBU). To optimize the distribution activity, it is necessary to design a scheduling system and its fuel delivery routes daily to every SPBU. The determination of routes and scheduling delivery of fuel to the SPBU can be modeled as a Petrol Station Replenishment Problem (PSRP) with the multi-depot, multi-product, time windows and split deliveries, which in this study will be completed by the Tabu Search algorithm (TS). This study was conducted in the area of Bandung, the capital of West Java province, which is a big city and the neighboring city of Jakarta, the capital city of Indonesia. By using the fuel delivery data for one day, the results showed a decrease of 16.38% of the distance of the route compared to the current conditions, which impacted on the reduction of distribution costs and decrease the number of total trips by 5.22% and 3.83%.

Strategic network design of Java Island fuel supply with production-transportation solution

This study aims to find more efficient supply network, from refineries / imports to fuel terminal, which still uses the Tanker, Tank Trucks or Rail Tank Wagon with an alternative pipeline that are considered more efficient than other transport modes, as well as gaining pipeline transportation network optimization analysis tailored to the capabilities/ capacity of refinery production and capacity of the pipe mode. With the complexity of the number of 3 point sources of supply, 19 destination of terminal, 4 kinds of products and 4 types of transport modes, transport-production model modified by adding multi-modal transport and investment costs of new pipeline. Then coded in Lingo program which adopts Branch & Bound technique and input the processed data in order to obtain an optimal distribution pattern produced the lowest distribution costs. This B&B solution was also compared with SCO solution which is a metaheuristic method. The results of this study lead to the development of new modes of pipeline connections in amount of 4 alternatives, generated from the optimal solution, but still potentially earned savings of about IDR 1 Trillion per year from cost-efficiency of product procurement and transportation costs.