scispace - formally typeset
Search or ask a question

Showing papers in "Industrial Engineering and Management Systems in 2021"


Journal ArticleDOI
TL;DR: The results of data processing indicate that the recommended office facility layout recommendations that are in accordance with the level of closeness between departments based on PST, ARC and ARD is the first recommendation, and with reduced total distance traveled in the first layout of the proposed office facilities, thelevel of employee productivity and work efficiency will increase.
Abstract: The electrical equipment industry in Indonesia is one of the factors that triggers economic growth. Labor productivity determines the success of project implementation. The non-optimal facility layout is one of the factors that triggers low labor productivity. The non-optimal facility layout affects the flow of information in the company. This study examines the problem of non-optimal office facility layout with TSP, ARC and ARD methods at PT. Duta Wijaya Elektrindo Engineering (PT. DWP). The method stated above can minimize the total distance traveled between office facilities so that work efficiency and workforce productivity may increase. The results of data processing indicate that the recommended office facility layout recommendations that are in accordance with the level of closeness between departments based on PST, ARC and ARD is the first recommendation. With reduced total distance traveled in the first layout of the proposed office facilities, the level of employee productivity and work efficiency will increase. The smallest total distance traveled on the proposed layout of the office facilities is 244.22 meters with a reduction of 36% from the initial conditions.

15 citations




Journal ArticleDOI
TL;DR: The results show that the performance of each forecasting method varies, depending on which demands of class ASKU are predicted, and the current forecasting method utilized by the company should be improved using the time series forecasting methods leading to the smallest error values for each class of A SKU.
Abstract: The company discussed in this paper is a national distributor firm that distributes FMCG products. The PPIC division in the company is responsible for forecasting the demand using the combination of the moving average method and intuition according to the interest of the company. However, the PPIC staff never measures the accuracy of their forecasting method. This research paper aims to evaluate the forecasting methods used to predict the demands of 12 classes of A SKU. Four-time series forecasting methods are particularly implemented, i.e., ARIMA, moving average (MA), double exponential smoothing (DES), and linear regression (RL). Forecasting using the ARIMA method is carried out by considering the stationarity of the average and variance of the historical data points. Forecasting using DES is carried out by using the optimal alpha and gamma values of the ARIMA method. The results show that the performance of each forecasting method varies, depending on which demands of class A SKU are predicted. Based on these results, the current forecasting method utilized by the company should be improved using the time series forecasting methods leading to the smallest error values for each class of A SKU.

3 citations



Journal ArticleDOI
TL;DR: 1. はじめに 近年,ECサイトを通じたオンラインでの商品購買に至る割合(Conversion Rate ; CVR)は通常,高々数%であることが知られている [2].
Abstract: 1. はじめに 近年,ECサイトを通じたオンラインでの商品購買は広く 一般消費者に受け入れられるようになり,市場規模が拡大し ている [1].ECサイト上では,ユーザの各ページの閲覧行動 の詳細なログが取得可能であるため,これらのデータを活用 したWebマーケティング技術の重要性が高まっている.一 方,ECサイトにおける購買に至る割合(Conversion Rate ; CVR)は通常,高々数%であることが知られている [2]. そのため,多くの ECサイトで CVRを改善するための施 策が必要とされている.例えば,ユーザの購買意欲を把握す ることで効果的なタイミングで施策を打ったり,購買につな がりやすいページを把握することでそのページにユーザを誘 導したりすることにより,CVRの向上が見込める. 一般に,ユーザは商品を購入する前に ECサイト上の様々 なページを閲覧することが多い.したがって,ユーザの思考 状態(購買意欲の有無や嗜好,ニーズ等)の変化は ECサ イトのページ遷移傾向に反映されていると考えられる.その ため,閲覧履歴データを分析することで,ユーザの購買意欲 を抽出できれば,適切な施策の一助になると考えられる.こ の様な分析においては,ユーザの思考状態に依存して各閲覧 ページが生成されたと仮定することは極めて自然であり,観 測不可能な思考状態を観測可能な閲覧ページから推測できる と考えられる.そこで本研究では,ユーザのページ遷移の背 後に潜在トピック(潜在クラス)を仮定するモデルを提案 する. 従来の潜在クラスモデルをページ遷移データに適用すると, 連続的に閲覧した一連のページ全体に同じ潜在トピックを仮 定するか,または各ページに対し毎回異なる潜在トピック を仮定するかのどちらかである.しかし,閲覧中にユーザの 思考状態が変化する可能性は十分に考えられるため,連続的 に閲覧した一連のページ全体に同じ潜在トピックを仮定する ことは好ましくない.一方で,ページごとにユーザの思考状 態が頻繁に変わることもほとんどないと考えられるため,毎 ページ異なる潜在トピックを仮定したトピックモデルの適用 も好ましくない. そこで本研究では, Hidden Topic Markov Models (HTMM)[3]をベースとし,リアルタイムに閲覧履歴データ の分析を可能にした購買行動分析モデルを提案する.HTMM は,前後関係を考慮しない Latent Dirichlet Allocation (LDA)[4]に隠れマルコフモデル(Hidden Markov Chain model ; HMM)の考え方を援用したモデルである.HTMM は文書に対して複数の潜在トピックを想定するが,LDAと は異なり,文書中の文単位での潜在トピックを考える.す なわち,各文中の単語は同じ潜在トピックに存し,かつ連続 する文は同じ潜在トピックを持つ可能性が高いことを仮定す る.このため,文書を同じ統計的特徴を持つ複数の文群に分 割することができる.したがって HTMMを閲覧履歴デー タに援用し,ページ単位での潜在トピックの推定を行うこと で,閲覧履歴を同じ特徴を持つ複数の群に分割することがで きる.そのうえで実際の閲覧履歴データから得られる購買実 績を併用することで,各潜在トピックの購買確率を求めるこ とができる.提案モデルにより,ユーザの購買意思をリアル タイムで予測することが可能になると考えられる.加えて, 各ユーザの潜在トピックの変化点を検出することが可能にな り,ユーザの思考が変化する際にどのようなページ遷移が起 こっているかを理解することができる.本研究では,実閲覧 履歴データを用いて分析し,提案手法の有効性を検証する. 2. 関連研究 2.1. 隠れマルコフモデル Hidden Markov Model(HMM)は,観測不可能なマ ルコフ過程とその各状態に依存して生成されるシンボルの組 み合わせによって,シンボルの系列を表現するモデルである. HMMの対象となる系列データは,複数の状態を持ち,それ らの状態がある遷移確率により遷移するマルコフモデルと仮 定される.これを一次マルコフ性という. 2.2. Hidden Topic Markov Model Hidden Topic Markov Model(HTMM)は,LDAに HMMの考え方を援用した,文書生成モデルである.図 1に HTMMのグラフィカルモデルを示す.

2 citations



Journal ArticleDOI
TL;DR: In this article, a preliminary study aims to compare production capacity at the family business of bricks at Sokaraja District, Banyumas Regency, Java, with three bricks SMEs for production capacity.
Abstract: This preliminary study aims to compare production capacity at the family business of bricks at Sokaraja District, Banyumas Regency. The majority of the businesses are done traditionally with family business managerial-based norms. The method used is the cost estimation analysis and the Analysis of Variance (ANOVA) method to calculate the statistical significance of production among SMEs. In addition, based on statistical analysis there are significant differences among the three bricks SMEs for production capacity. The implications of these calculations can be used as an initial step to consider in general determining the logistics conditions, sales strategies, and the role of succession in the family business. In practice, increasing production capacity, reducing production costs, and product innovation can be achieved if the involvement of family members were optimal to ensure and enhance the business’s sustainability.

1 citations


Journal ArticleDOI
TL;DR: In this article, a multi-objective mixed-integer mathematical model is presented, including the objectives of minimizing costs and products with delays and maximizing the total purchase value for garment supply chain.
Abstract: The garment supply chain is one of the most common supply chains in the world. In this supply chain, quality and cost are the most important factors that are strongly related to the selection of suppliers and the allocation of orders to them. Accordingly, the purpose of this paper is to integrate decisions for supplier selection, order allocation, and multi- source, multi-mode, multi-product shipping plans with consideration of discounts under uncertainty. For this purpose, a multi-objective mixed-integer mathematical model is presented, including the objectives of minimizing costs and products with delays and maximizing the total purchase value. In this mathematical model, the policy of purchasing materials and determining the number and type of transport equipment are specified. To solve this mathematical model, a goal-flexible programming approach with a utility function is presented. In the solution algorithm, a new possibility-flexible programming method has been developed to deal with the uncertainties in the model, which is based on the expected value method and chance constraint. Finally, using a numerical problem, the establishment of the above model in the garment supply chain is investigated. As indicated by the outcomes, the proposed model was touchy to certain boundaries, including blended leaders’ mentality, a boundary identified with fluffy imperatives, and the degree of certainty characterized by the chief for not exactly equivalent limitations.