This thesis is to investigate the general characteristics of MIS and provides a definite data base for computerization of the division operations based on the analysis and design of personnel business. Therefore, through the analysis of DFD (Data Flow Diagram), DD (Data Dictionary) and SD (System Dictionary), for designing of system, and the analysis of the requirement information, the relationships of business and the process of operations, for the current status analysis of division business, the computerized alternatives corresponding to a choice of a piece of information system were decided. The priority of a developing operations was also determined. A point of view is not so much the study about the division operations so far. This paper will be very helpful and will be achieved through the computerization of MIS.

Hangout programming language, written in Hangout and having syntax of Korean, can enhance readability very much for its naturalness, and is suited for Hangout data processing. But there exist several problems in implementing it due to the peculiarities of Korean such as verb position, variation of termination. This paper proposes a solution for implementing Hangeul high level language and introduces a Hangout programming language HANSFOR (Hangout Structured Fortran), implemented according to the proposed method. HANSFOR satisfies such requirements of Hangout high level language as mentioned above and can handle Korean and Chinese character data efficiently. Especially, though it is based on Fortran, it is free-formatted and provides structured programming concept to compensate for defects of original Fortran. It is successfully implemented on NEC S/100.

This paper deals with a transportation problem with two objectives, that is, transportation cost and vital-route shipment. In the transportation problem minimizing cost is sometimes conflicting minimizing vital-route shipment. This paper develops an algorithm finding paneto-optimal solutions in minimizing vital-route shipment and cost, and also presents theoretical backgrounds for developing an algorithm.

Military personnel encounter a variety of noise environments. During exercises, high intensity noise levels are often encountered. Twenty-four subjects were required to respond to symbols presented under two levels of task difficulty, two levels of presentation rate, two levels of display orientation, and three levels of noise intensity. The purpose of the experiment was to determine whether noise intensity and display orientation had any effect on a short-term memory task. Results showed that continuous white noise at intensity levels of 30, 85, and 105 db had no effect on the shortterm memory task. Presentation rate and task difficulty demonstrated a significant relationship with task performance as did their two-way interaction. This two-way interaction between presentation rate and task difficulty exhibited a different pattern for the two levels of display orientation.

A block replacement policy using items with different reliability is discussed. We divide system unit failure modes into two modes and use less reliable unit when operating unit fails near the planned preventive replacement time. In this policy, item A has two failure modes. Mode-1 failure is removed by minimal repair, mode-2 failure by replacement. If mode-2 failure of item A happens in (0, $T-{\delta}$), failure item A is replaced by new item A. If mode-2 failure of item A happens in ($T-{\delta}$, T), failure item A is replaced by new item B. Item B should be cheaper and less durable than item A. Under this policy, we determine the preventive replacement interval $T^{*}$ and the interval ${\delta}^{*}$ of item B replacement which minimize the cost rate per unit time.

Consider an m machine flow shop with blocking. The processing time of job j,j=1,..., n on each one of the m machines is equal to the same random variable $X_j$ and is distributed according to $F_i$. We assume that the processing times are stochastically ordered, i.e., $F_{1_{-st}}{<}F_{2_{st}}{<}cdots_{-st}{<}F_n$. We show that the sequence 1,3,5,...,n-1,n,n-2,...,6,4,2 when n is even and sequence 1,3,5,...,n-2,n,n-1 ... 6,4,2 when n is odd minimizes the expected makespan and that the sequence 1,...,n minimizes the expected flow time.