One of the problems in supply chain management (SCM) is the bullwhip effect, a phenomenon that amplifies fluctuations in the quantity purchased or ordered. Fluctuations in the quantity ordered by customers are small, fluctuations in the quantity ordered by retailers to wholesalers are large, and fluctuations in the quantity ordered by wholesalers to manufacturers are even larger. The amplification of demand fluctuation makes supply-demand management and production management difficult, making it an important issue for copier manufacturers and their component manufacturers located upstream in the supply. The bullwhip effect is supposedly affected by factors such as inaccurate demand forecasting, long production and logistics lead times, and minimum order quantities. This study developed a demand forecasting method for production control by clarifying and improving the factors that make it difficult to accurately forecast demand for products sold by copier manufacturers. Furthermore, we developed a demand forecasting method based on a voting system using a collective intelligence mechanism. The proposed method was devised to overcome problems such as the “beauty voting” problem in existing forecasting methods (prediction markets) using collective intelligence. A comparison of six months of data during the study period showed the amount of inventory by approximately 80% and reduced the shortages to zero for the product under consideration.

We consider the characteristics of a consignment-based contract with an exogenous retail price in a two-level supply chain that consists of a retailer and a manufacturer. The retailer is in a newsvendor setting, and the market demand is sensitive to the retailer’s sales effort. Before the selling season, the manufacturer offers the retailer a wholesale price. The retailer then simultaneously determines 1) the prospective degree of sales effort for the season and 2) the order quantity for the manufacturer. At the end of the season, the retailer remits a wholesale price to the manufacturer for each item sold. Unsold leftover items, if any, are collected and salvaged by the manufacturer at the manufacturer’s cost. This type of contract can be seen in the Japanese fashion industry. We will focus on how a unit wholesale price set by the manufacturer affects the profit of both the retailer and the manufacturer. The research questions we seek to answer are as follows: As the retailer seeks to maximise his own profit, 1) how does the wholesale price determined by the manufacturer affect the retailer’s response, and 2) how does the retailer’s response affect the manufacturer’s and the channel profits? A threshold wholesale price, which determines the retailer’s optimal reaction to the wholesale price, exists, and the market sensitivity to the retailer’s sales effort has a significant impact on both the retailer’s and the manufacturer’s performances.

Smart factory became one hot topic both in research and practice. Technologies supporting smart factory can be divided into four categories, i.e., automation, information and communication, intelligence and optimization. These four technologies can be grouped into hard technologies and soft technologies. Soft technologies mainly supported by management science and industrial engineering. The soft technology which mainly used in smart factory, we briefly call the management technology, which can maximize utilization of the hard technologies, therefore it can be seen as the technology on technology. Assuming that a company constructed fully automated factory, however, the automated machines breakdown frequently, many automated machines face to starvation (idling), the automation technology could not contribute competition power of the company. As the same reason, information and communication technology can collect and transfer huge amount of data, however, if the data do not processed on time for supporting correct decision-making, then the value of data will be zero because data collected in factory is perishable. In this research, we focus on the management technology, temping to propose a method to solve major problem which makes gaps between periodic production plan and daily production scheduling. Current production scheduling problem assumes the periodic production plan is given, which may generated by so called MRP (material requirement planning) logic using BOM (bill of material) information. The problems here are that the periodic production planning neither consider precedence constraint of each job, nor consider dependency of processing time with production lot-size. Furthermore, MRP logic require a fixed leadtime as a parameter, therefore, the daily production schedule generated using MRP logic frequently results in infeasible. We propose an integrated production planning model, generating the periodic production plan incorporating daily production scheduling same time. In the proposed model, we introduce varying processing time which depends on lotsize, introduce two different lot-size for each job, and find optimal combination of the lotsize under the given capacity constraint during the planning horizon, while production load is calculated based on the optimal scheduling problem for each combination. Through numerical examination, we justify the proposed model, and analyze trade-offbetween cost increase and operation time of the manufacturing system.