Sadamichi Mitsumori

Optimum Production Scheduling of Multicommodity in Flow Line

The applications of computer control in production factories have been extended from the control of mass and energy to the production control. The control at this level comprises the formulation and alteration of work schedules in accordance with the progress of work, and it may be called the control of information. This paper deals with the flow line which is the most fundamental production line in the factory and proposes and verifies the optimal scheduling method. In the case of producing multicommodities by a flow line, the optimum schedule is the one which satisfies the demand for the respective products and minimizes the changeover loss. The branch-and-bound method is used to obtain an optimum schedule. The geometrical characteristics of the region in which the feasible schedules exist are used for calculating the lower bound of the objective function for the subset of feasible schedules.

Corporate information systems and information technologies

A corporate activities business-cycle model composed of PLAN, DO, and SEE activities is described. This model is constructed on two layers: an organizational layer and an individual layer. Regularity of both data structures and algorithms is defined from program production aspects as follows: if the data structures and algorithms of a program can be defined prior to all computations, they are regular, otherwise they are irregular. It is shown that unique information technologies are required to develop each system. Several information technologies are proposed for each information system: real-time database maintenance for DO information systems, a data supply system for SEE information systems, an automatic index generator and a virtual office environment for PLAN information systems, and private data management and automatic terminal operation for PLAN and SEE information systems at an individual layer under organizational DO activities. >

Optimal Scheduling for Load Balance of Two-Machine Production Lines

A scheduling method for balancing work loads between a machine and the one succeeding it in a two-machine production line so as to maximize the overall production speed is proposed in this paper. The case where the production speeds are different from each other both for machines and for commodities is treated. The objective is to minimize the changeover loss under the constraints that 1) the due date must be met, 2) the queue length of commodity items must be shorter than the maximum space between the two machines, and 3) no idle time is acceptable for either machine. Both the lot size of each commodity for each production period and their production sequence is determined by the proposed method. The scheduling algorithm is a branch-and-bound procedure. In the conventional flow shop scheduling problem, the lot size of each commodity is given beforehand considering the inventory loss of the final products and the changeover loss of the production line. Only the flow sequence is determined there. Therefore, it can not perfectly balance the work loads between machines. Computational experience with the proposed method indicates that it is capable of solving problems involving 20 commodities over a planning horizon of 300 time periods. These are practical scheduling problems with a planning horizon of one week for a home appliance production line. These computation times are below one minute (using the HITAC M-180 operating at three million instructions per second (MIPS)).

A Resource Allocation Method and an Identification Method for Discrete Systems

Abstract This nor Loposea. a nov construction for resource allocation problem(RA-Problem). A new PA-problem concidering both job sequence and Material flow in given and solved. Job sequence and material flow are expressed as a network model. This paper also proposes an identification method for PERT arrow-digram, which is the most simplified network model. The proposed identification method is heuristic, but it can evaluate the validity of identified values. An example is also given where the proposed methods were applied in the designing of n warehouse system.