마이크로 컴퓨터를 이용한 多品種 少量生産 시스템의 전산화 事例

宋漢植 東亞大學校經營問題硏究所 1984 經營論叢 Vol.7 No.-

This paper showed the application of Micrcomputers(IBM-PC compatibles) to (ⅰ) the order entry and delivery data processings, (ⅱ) stock items keeping and inquiry, and (ⅲ) grouping the ordered items by their specifications or by their raw materials requirements in the custom products job shop or batch process systems. TAEKWANG BEND IND. CO. which produces bends products were the real example of this applocation study. Bends are specified by their shape(item name), types, schedules, sizes and raw materials, and ordeed items are specified by their custom name, order no, p. o. no. duedates,, and other remarks. In this company, above (ⅰ), (ⅱ), and (ⅲ) were serious problems because of its diversity of products and of its complexity of order entry and delivery processings. Thus it has been tried to develop the EDP systems for those problems by using an IBM-PC compatible computers. EDP system development was through the several stages: system analysis, system design, data base system design, program development and installment. For the easiness of system development, dBASE Ⅱ from Ashton-Tate was used. dBASE Ⅱ is relational data base management systems which is easy and powerful in system design and program development. Program was developed as menu driven system so as to perform the task by choosing proper options. Computer system was made up of main body, 10 mega byte hard disk, and printer as microcomputer system. Now days, because of continuous falling of hardware prices, it is better to computerize the daily transactions and to support the plan and control of the custom-products job shop or batch systems by using the micro level computers.

제조업에서의 노동의 양극화

송한식 東亞大學校 1998 東亞論叢 Vol.35 No.-

This study concerns the dichotomies in manufacturing work: industrial manufacturing work and postindustrial manufacturing work. Industrial manufacturing work consists of making, moving, inspecting, loading and unloading. However postindustrial manufacturing work includes flew product design, research and development, and programming computer aided manufacturing machines. Four industrial manufacturing systems were visited as for case studies to analyze the dichotomies of work: pin manufacturing system, embroidering manufacturing, shoe making, and casting manufacturing system. The dichotomies in work were investigated in the respect of input/output of work, necessary work capabilities. and work duplication, work substitution. Also several issues from the dichotomies of work were distinguished such as social and economical issues, and individual, company, and governmental issues were investigated.

OR/MS의 스프레드시트 모형화의 장단점과 문제점 : Its Advantages and Disadvantages

송한식 동아대학교 경영대학 2001 經營硏究 Vol.- No.14

This study figured out the advantages and disadvantages of modeling with spreadsheet tools. It compared with LINDO with reasonable size of LP problems. There were 6 advantages. 1. Problem description and problem modeling can be done on the same spreadsheet. But it can not be done on LINDO. 2. Spreadsheet modeling helps communications among the users in conrcern by using the same tool. 3. Spreadsheet modeling helps understanding and expression of the problem. 4. Spreadsheet modeling is easy to accomodate multistage implicit relations in problem. 5. Secondary analysis on the sensitivity reports can be done on the same spreadsheet. 6. It can use macros for further analysis and model utilization. However, spreadsheet model is hard to verify its modeling. Mistranslation of the Solver terminologies into Korean causes serious differences between text books and spreadsheet modeling words such as 'Target Cell' and 'Objective Cell'.

不良損失을 고려한 多段階 로트生産의 生産스케줄 樹立에 관한 硏究 : 신발類 製造業體를 實例로 in a Case of Shoes Production

宋漢植 東亞大學校 大學院 1984 大學院論文集 Vol.9 No.-

The competitive strategy of an organization is represented by its policies with respect to the marketplace and its competitors; four major competitive weapons are cost, quality, availability and flexibility. For intermittent system, when the quality and flexibility are almost equal among competitors, cost and availability are critical factors to have advantage over other competitors. Intermittent system has two important characteristics regarding availability and cost. One is learning curve phenomenon and the other is scrap factor in production process. So, in this paper a model which finds the efficiency of system is proposed to represent competitive power from availability and cost which are under the influence of learning curve phenomenon and scrap factor. Especially, scrap factor is the cause of delay of order delivery and increase of cost and resulted in the deterioration of productivity. On the other hand learning, curve phenomenon in production process of one kind of item reduces the production time and costs. So, this paper has the idea to represent opportunity cost of waiting time of order and defective losses into profitability of system under certain scheduling rule. Besides profitability, average number of waiting orders and average waiting time of an order are considered as systems efficiency. To find profitability as profit per systems hour, the difficulty lies in to find out the waiting time of orders. Two methods are considered. One is by analytic queueing M/G/m model or Level Cross Analysis model. The other is Simulation model. Queueing model is difficult to accommodate practical real requirements into theoretical model and to expand mathematically. So, Simulation model is constructed. The imput variables are learning curve percentage (R), defective rate (p), set-up time (UP), rate of return (r), prices (P), raw material and labor costs (C), and feedback probability (d) to complete scrap deficiency. When these input variables are given, the simulation model finds the profitability, idle time of facilities, average number of waiting order (N), average waiting time (W),etc. Through numerical example of shoes production case, it was found out that scrap factor has severe influence on profitability than expected, and average number of waiting order was greatly influenced by systems utilization factor (ρ) and average waiting time was consistently stabled through the experiment. In addition, if we have practical past history of order arrival and order size, these data can be used to generate order arrival data, we can get profitability and other important systems characteristics which give some important information about the operation of intermittent productive system such as order processing rule or scheduling priority rule.

意思決定樹 作成 問題의 分析

宋漢植 東亞大學校 大學院 1989 大學院論文集 Vol.14 No.-

Constructing an optimal decision tree for a given decision table is an NP-Complete problem. A search algorithm for general graph searching problem has been applied to the decision tree construction problem. An estimation function h in the A( )search has been defined by using entropy of the set of unknown elements. The difficulties of the A( )search in decision tree construction have been found : the correlation of testing variables, and the fact that entropy reduction by testing a variable may be less than 1.0 The idea of new estimation function has been introduced to overcome the NP-Complete obstacles : the estimation function must reflect both the correlations among variables and the entropies of test variables. The conditions of estimation function are specified in order for the A( )search to be an"efficient" algorithm. To find out such a new estimation function is claimed to be a clue to the answer to the question of NP=P.

GMDH를 이용한 자동모형화에 관한 연구

송한식,엄홍섭,안기범 東亞大學校 經營問題硏究所 1996 經營論叢 Vol.17 No.-

SUMMARYWe now that system's modeling is to find the functional formulation of system's transformation process from input variables to output value. In this causal relationship, we assume that transformation process has very stable and predictable function and we can only determine parameters of the function.But in this system's modeling process, we bring up two problems. One is, in much cases, we have no conviction that really they have functional relationship. The others is, even if they have functional relationship we can not know the structure of functional equation.To solve this problem, we need a new method which could build functional formulation of itself. It is called self-organizing methods of modeling. GMDH (Group Method of Data Handling) is one of these self-organizing methods.In this paper, the GMDH is used to build a model in some cases.The GMDH algorithm, which was introduced and developed by Ivakhnenko, constructs high-order functional models for complex systems. GMDH self-organized the model from a simple one to one of optimal complexity by a methodology using the process of survival-of-the-fillest principle to determine which new equations to live and which equations to die. This new generation of equations will then be better suited to describe our system then the original this GMDH algorithm has thee steps in detail below. step 1 : constructing new variables Z1, Z2, ......Zm(m-1)/2 step 2 : screening out the less effective variables step 3 : test for optimality We developed this GMDH algorithm to a computer program. We applied GMDH to estimate the function which gives the value of degree of freedom by using Chai square(x2) distribution probability values and time series analysis of stock price estimation. In this trial, we used some previously-known data and the estimation ability of GMDH was good.In conclusion, the GMDH is excellent to build model to estimate dependent variable without specifying functional formulation in system modeling. So, in future study, we can apply this GMDH algorithm to develop the structure of neural network in forecasting and the area of medical science, economic model, etc.

정신분열병 환자의 반응시간

최형식,정범석,김한규,송현주,이재원,김창윤 大韓神經精神醫學會 2000 신경정신의학 Vol.39 No.5

연구배경: 주의력은 인지적인 정보처리과정(information processing)의 가장 기본적인 단계이다. 반응시간 (reaction time)은 주의력을 가장 잘 반영하는 검사 중 하나로서, 자극이 제시된 후 반응이 시작되기까지 반응 여부를 결정하는데 걸리는 시간인 판단 시간(decision time)과 반응이 시작되면서 목표에 도달할 때까지 시간인 운동시간(motor time)으로 나누어 볼 수 있다. 기존의 연구결과로는, 정신분열병 환자군에서 반응시간의 수행이 떨어져 있다는 보고는 많이 되어왔으나 차이가 없다는 보고도 있으며, 판단시간과 운동시간 중 어느 부분에 장애가 있는지 명확하지 않다. 또한 반응시간 측정값은 질병의 진단적 유형(diagnostic subtype), 증상, 경과 등의 영향을 받으며 개체간의 변동(inter-individual variation)이 크고, 같은 피험자라도 측정시점이 다를 때마다 차이를 보이는 개체내 변동(intra-individual variation)이 커서 검사결과마다 차이를 보이고 있는 상태이다. 목적: 환자군은 개체간 변동이 크고 다른 임상적 요소에 의해 영향받는 점을 고려하여, 임상적으로 안정적인 경과에 있는 정신분열병 환자만을 대상으로 정상인 군과 반응시간의 차이를 비교하고자 하였다. 또한 반응시간을 판단시간과 운동시간으로 나누어 측정하여 각각을 비교분석 하고자 하였고, 개체내 변동이 큰점을 고려하여 25회 반복 측정을 실시하고 이에 따른 변동의 경향성을 조사하고자 하였다. 방법: DSM-Ⅳ에 의해 '정신분열병'으로 진단된 환자로서, 급성기 이후 항정신병약물을 유지하고 있으며BPRS(brief psychiatric rating scale) 30점 이하의 임상적으로 안정적인 상태가 3개월 이상 유지된10명이 환자군으로 포함되었고, 정상인은 자원자 10명이 포함되었다. 검사도구는 비엔나 테스트 시스템 반응유닛(Reation Unit) 파라미터 S8을 이용하였다. 총 6가지의 자극이 무작위로 1측정시기(session) 당 총 32회 주어지고, 피험자는 이중 특정 자극에 8번만 반응하도록 되어 있다. 시작 전 1회 연습 후 총 25측정시기를 5분 간격으로 반복하였다. 반응 시간을 판단시간과 운동시간의 합으로 정의하고 세 가지 모두 컴퓨터를 이용하여 자동 측정되었다. 반응·판단·운동 시간의 비교에는 반복 측정을 고려한 분산분석(repeated measures analysis of variance)을 사용하였다. 결과: 환자군은 정상인에 비하여 유의하게 반응시간의 수행속도가 저하되어 있었다(p<0.01). 판단시간과 운동시간으로 나누어 볼 때, 환자군에서 판단시간이 유의하게 느렸고(p<0.01), 운동시간은 수행저하의 경향을 보였으나 통계적으로 유의하지는 않았다. 25회 반복측정시 양군 모두 측정시기에 따른 유의한 변화를 보였고(p<0.01), 환자군의 반응시간은 개체간의 차이뿐 아니라 개체내 변동이 큰 경향을 보였다. 옳게 반응한 횟수(correct response)는 환자군에서 저하되는 경향을 보였으나 통계적으로 유의하지는 않았다. 고찰: 정신분열병 환자군에서 판단시간이 유의하게 저하된 것은, 반응시간의 저하가 운동요소 보다는 인지적 요소에 의한 것임을 시사하는 것으로, 정보처리과정의 기본적인 단계인 주의력 장애를 보여주는 것으로 생각된다. 비교적 안정적이고 임상적인 특성이 동일한 환자군에서도 반응시간이 대조군에 비해 큰 개체간 차이와 개체내 변동성의 경향을 보였으며, 이는 단순한 측정오차라기 보다는 정신분열병 환자군의 인지기능의 특성을 반영하는 것으로 생각된다. 중심단어:정신분열병·반응시간·판단시간·운동시간·개체간 변동·개체내 변동. The reaction time(RT) has been known to reflect attention that controls the flow of information processing. Extensive research has demonstrated cognitive impairment in schizophr-enia subjects using RT taske. However, little work has been done examining the relative contribution of DT(decision time) and MT(motor time) to slowed RT in schizophrenics. Also, recent investigators have observed that schizophrenic patients exhibit larger intra-individual variability in RT than do normal comparison subjects. The purpose of this study, using multi-stimulus convergent RT task, was to explore the speed of RT, relative contribution of decision time(DT) and motor time(MT) to slowed RT, overall sequential profile in 25 repeated-time measurements in 10 schizophrenic out-patients and 10 normal control subjects. Overall reaction time and decision time were slower in schizophrenic subjects than in normal controls. The motor time was not shown to be significantly different between the two groups with 0.05 significance level, although there was some trend indicating schizophrenic subjects were slower consistently in repreated measurements over time. These results suggested that the slower reaction time in schizophrenic subjects was mostly determined by cognitive component, decision time rather than motor time. Sequential profile of repeated measurements showed greater intraindividual and interindividual variations in schizphrenics than in normal controls. These results indicate that high variabilities are not merely measurement errors but characteristic of schizophrenic psychopathology. KEY WORDS: Schizophrenia·Reaction time·Decision time·Motor time·Inter-individual variation·Intra-individual variation.

신발 제조업에의 린 생산방식 도입

윤재홍,송한식,권치명,정문상 東亞大學校 經營問題硏究所 1994 經營論叢 Vol.15 No.-

This paper is concerned with a lean production system to improve the productivity of shoes manufacturing industries. For this purpose, this research first reviews the environmental characteristics of shoes manufacturing industries and invetigates the current problems of shoes production systems. Shoes manufacturing process roughly consists of three process lines shoes uppers line, shoes soles line, and assembling line of shoes uppers and soles. Two major problems in shoes production system are a large amount of work-in-process(WIP) inventories, and the low rate of matching shoes uppers and soles in an assemblying line. Most of shoes manufacturing companies have more than a 3-days WIP inventories in every process line to avoid the blocking the production. Particularly the assemblying line has 5-7days buffers of shoes uppers and soles to maximize the productivity of manufacturing final shoes, however, the matching rate of shoes uppers and soles in the assemblying line is about 50%. These problems are from a low efficiency in inventory control mainly due to (1) too many order specification of sizes, colours, types and different due dates, and (2) a large amount of WIP inventories. To reslove above mentioned problems, we sugggest a lean production system of pull-method for reduction of WIP inventories and improvement of matching rate of shoes uppers and soles in assemblying line. Compared to a push production system(current production system), a pull production system may have a number of adventages in controlling inventories and balanced match of the sole parts and uppers parts when it applied to shoes manufacturing process. As an alternative to control the pull production method, we design a Kaban for production and inventory control. Also we design the Kaban circulations model for the production lines and inventory stores, and a shoes manufacturing information system to incorporate the Kaban system. We expect that the effective adoption of the pull production system to shoes manufacturing system will show better results in improving the matching rate, and reducing the WIP inventories and production lead time. These effects may also give more flexible production planning and raw matiarial procurement planning. We finally consider this research would be helpful in designing the lean shoes production system

미만성 경화 아형 갑상선 유두암 1예

최정민,이송주,김부경,최영식,박요한,김정훈,허방 고신대학교 의과대학 2009 고신대학교 의과대학 학술지 Vol.24 No.1

The diffuse sclerosing variant (DSV) is a rare malignant tumour that accounts for 0.7-5.3% of all papillary thyroid cancers. DSV tends to occur at a younger age and has a higher incidence of cervical lymph node metastases, compared with papillary thyroid carcinoma. A 62-year-old female patient diagnosed thyroid nodule with diffuse and mild increased 18F-deoxy-glucose (FDG) on FDG positron emission tomography (PET)-computed tomogram (CT) scan. The ultrasound images showed typical findings of DSV such as diffusely enlarged thyroid lobes and diffusely scattered microcalcification. We performed this patient ultrasonography guided fine needle aspiration and cytologic analysis revealed papillary carcinoma. Total thyroidectomy with central compartment node dissection and lateral lymph node dissection was performed and finally diagnosed DVS. The authors report this case with literature review.

표재성 방광암에서 Bacillus Calmette-Guerin 방광 내 주입요법 : 장기적 추적결과 Long-term Follow up Results

김용웅,왕세환,육승모,한동석,노안식,나용길,설종구 충남대학교 의학연구소 2003 충남의대잡지 Vol.30 No.2

To evaluated the results of long-term follow up of superficial bladder cancer patients who were treated with transurethral resection and intravesical BCG therapy. Between April 1989 and December 1997, 65 patients with superficial bladder cancer were treated with TUR-B(Transurethral resection of bladder tumor) and intravesical BCG instillation, and follow-up was performed for 5 years longer, and the mean follow-up duration was 82 months. Of 65 patients, pathologic stage were Ta in 17 cases and T1 in 48 cases and grade were G Ⅰ in 22 cases, G Ⅱ in 35 cases and G Ⅲ in 8 cases. The tumor recurred in 22(33.8%) at the median of 21.8 months and the tumor progressed in 9(13.8%) at the median of 20.8 months. Most of the recurrences and progression were within 3 year(838 and 77%). Progression is more likely in higher grade(G Ⅱ, G Ⅲ) and higher stage(T1). The complete response rate after first and second courses of BCG therapy were 53% and 73% and the 5 and 10 year disease specific survival rates were 89.5 and 78% respectively. The complete response rate after first and second courses of BCG therapywere 53% and 73%. Most of recurrences and progressions were within 3 years of the BCG therapy, but some cases occured after 5 years. Therefore, close long-term follow up is needed.