This paper discusses with the integration of order bidding and manufacturing planning decisions that is, considering contingent orders and multiple production resources. The main objective of this research is to integrate decisions in the internal supply chain of a make-to-order (MTO) manufacturer. Our paper is positioned as subsequent research to Cakravastia et al. (2002), Cakravastia and Nakamura (2002) and Cakravastia and Takahashi (2002), which previously addressed the integration of manufacturing and procurement activities. In this paper, we focuson marketing and manufacturing integration. Since an order for an MTO manufacturer usually comes from a winning bid, we propose a model and technique to simultaneously determine the price and lead-time bidding decision for a new order. We take contingent orders into consideration because these orders, with some degree of probability, will arrive at the shop floor, causing workloads to rise at various facilities; thus affecting production control, scheduling, and hence cost. We extend the bidding decision model of Easton and Moodie (1999) for the case of multiple resources and multiple jobs for each order. As the fundamental objective of a bid is to provide profit to the manufacturer, instead of just winning the order, we integrate the bidding decision with the manufacturing planning decision in preparing the bid for a new order. This decision is responsible for defining : (i) job schedule, and (ii) fabrication lot size. By this integration, the consequences of each possible backlog scenario for manufacturing activities can be determined. The result shows that the proposed technique can be used to optimize the bidding decision and avoid over optimistic bids that would lead to the "winner curse" phenomenon.

View full abstract

This paper discusses and proposes a systematized product development process based on an idea that product life is made of six structurally very similar engineering effort sub-cycles. The new product development process is a major part of multiple sub-cycles within the total product life. The development program starts from the initial idea stage and extends to the ending stage at the discontinuation of product support. At the early stage of this study, the original task node diagram was viewed as a large network diagram of approximately 1,000 nodes requiring the equivalent of 30 feet of connected letter-size paper. For the purpose of illustration, this continuous paper was coiled on to a spiral surface with the proper diameter to line-up the starting points of six sub-cycles, namely : Idea, Breadboard, Prototype, Pre-production, Manufacturing and Product maintenance. In many cases of product development, contract programs may cover only two to three sub-cycles each, such as the Bread-board, Prototype and Pre-production sub-cycles. Each of these six sub-cycles resembles one another with a nearly similar pattern. In general, the total of the smallest task stages observed within one sub-cycle was 18. Three task stages were grouped together to make six groups in one cycle. On one sheet of 8 1/2×11-in. paper, these three stages were shown together. Six pages hold one sub-cycle, and a total of six sub-cycles covers a complete product lifecycle totaling 36 sheets. In this proposal, several applications and evaluations of the concept are discussed. The discussions include : Node diagrams to serve as program status reporting media, progress control indicator, inter-departmental communication tool, accurate budgeting and cost estimating tool, and program control tool for multiple parallel product development operations.

View full abstract

This paper discusses an optimal control problem for a failure-prone multi-stage production line that minimizes expected total cost per unit time. The problem is formulated as an undiscounted Markov decision process (UMDP), but it is difficult to solve the problem exactly due to the curse of dimensionality. In recent years, however, several algorithms in the field of reinforcement learning, or neuro-dynamic programming (NDP), have been devised to overcome the curse of dimensionality. In this paper, the simulation-based modified policy iteration method (SBMPIM) is proposed as a new algorithm of NDP. The SBMPIM adopts simulation in the value approximation routine of the modified policy iteration method (MPIM), which is an algorithm of UMDP applicable to relatively large-scale problems. The SBMPIM and existing algorithms, such as SMART, RELAXED-SMART and SBPI, of NDP are numerically compared with an optimal control policy of a single-stage production line computed by MPIM. Moreover, the performance of JIT production lines with the optimal numbers of kanbans is numerically compared with that of optimal or near-optimal control policies computed by SBMPIM. It is shown that all algorithms except the SBMPIM of NDP fail to converge, and the expected total cost per unit time of JIT production lines with the optimal numbers of kanbans rises by more than 7 percent in comparison with the expected total cost per unit time computed by SBMPIM.

View full abstract

This paper proposes an improved SLA^* (search and learning A^* algorithm) for solving resource-constrained project scheduling problems. The objective is to reduce the amount of SLA^* backtracking. Resource utilization and number of parallel processing activities are two new criteria of the proposed algorithm. The proposed algorithm combines those three criteria : (1) resource utilization, (2) total heuristic estimate, and (3) number of parallel processing activities. The composite criteria are used at the front state selection of the SLA^* procedure. The performance of the proposed algorithm is analyzed with different problems. The problems are characterized by problem scale, network complexity, resource factor and resource tightness. Computational results show that the proposed algorithm reduces backtracking more than 50% on average compared to the original SLA^* giving optimal or near optimal solutions.

View full abstract

For a make-to-order production-inventory system with parts as its input, parts should be replenished giving consideration to the inventory cost of parts and back-logged cost of demand. In this paper, we discuss a discrete time make-to-order production-inventory system with a constant production capacity and stochastic demand. Parts used in the production process are supplied by a vendor with constant lead-time. The inventory of parts is controlled by a replenishing point system. A necessary and sufficient condition under which this system provides stationary distribution of the back-logged demand is derived. This stability condition is obtained by a method similar to the analysis of GI/G/1 queueing system. It is represented with the demand distribution for each period, the production capacity of the system, the lead-time of parts and the replenishing point of parts. The behavior of the back-logged demand is discussed for the case of deterministic demand. The probability generating function of back-logged demand in the equilibrium state is derived and the average total cost per period, average back-logged demand and average inventory of parts are expressed in terms of demand distribution, production capacity, lead-time of parts and replenishing point of parts. With these results, an algorithm is developed for determining the optimal replenishing point of parts. Some numerical examples are given to show its efficiency.

View full abstract

This paper proposes the development of a software simulator that allows users to build a supply chain model and analyze the logistics issues behind the model. The simulator consists of five kinds of participants and an e-marketplace for the supply chain. Each participant is implemented as a distributed object such that it runs concurrently and has the following capacities and policies : (1) end customers, (2) intermediaries including manufacturers, (3) parts suppliers, (4) electronic payment service providers, and (5) transportation servers. The e-marketplace defines the trade protocol for the workflow management and transaction analysis. The SCM simulator visualizes goods flow, money flow and information flow. This paper discusses the background of the proposal, the goal of simulator, and milestone, and technical issues for development.

View full abstract

A job consists of several processes, and the job process line describes the job process content, orders and paths. If these are improved, productivity is expected to increase. The process analysis method (PAM) concept is introduced in this paper. PAM uses spreadsheet software to improve job processes and support job implementation. PAM users input job process data, which is mainly recorded upon the execution of each job process. Data such as date and time of process execution, time required, code number of the person in charge, etc. is recorded. As a result, PAM users can analyze and evaluate the data, and redesign the job process and flow. Furthermore, the redesigned job process data is input and analyzed. This effort results in alternative redesign plans. In the latter half of this paper, the daily job execution support system on the spreadsheet for the staff in charge and manager is discussed. For the PAM method, the processes flow line is designed using an Excel spreadsheet. One process (activity) is described as a square box consisting of nine cells. In each cell of a process, various information for that particular process is provided. From the data on each process and the flow chart input into the sheet, a business process chart is automatically drafted by an Excel's Visual Basic Application program we developed. This system can be utilized with popular spreadsheet software; therefore, anyone who has the spreadsheet on their personal computer can be offered the PAM utility. This system intends to support both redesign work and work process execution.

View full abstract