• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2013.05a
• /
• pp.759-760
• /
• 2013

• Management Science and Financial Engineering
• /
• v.2 no.1
• /
• pp.19-51
• /
• 1996

We suggest a distributed framework for task assignment in the computer-controlled shop floor where each of the resource agents and part agents acts like an independent profit maker. The job allocation problem is formulated as a linear programming problem. The LP formulation is analyzed to provide a rationale for the distributed task assignment procedure. We suggest an auction based negotiation procedure including a price-based bid construction and a price revising mechanism. The performance of the suggested procedure is compared with those of an LP formulation and conventional dispatching procedures by simulation experiments.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.75.2-75
• /
• 2001

In this paper, we provide the technique of H$\infty$ controller design algorithm for input-saturated linear systems using a linear parameter varying(LPV) framework. The LPV controller with parameter dependent dynamic state feedback controller concept guarantees the asymtotic stability and H$\infty$ norm bound within prescribed level v using the saturation nonlinearity as scheduling parameters. Especially, the sufficient conditions for the existence of H$\infty$ controller are formulated in terms of linear matrix inequalities(LMIs) that can be solved very efficiently.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.17 no.3
• /
• pp.171-180
• /
• 1992

This paper proposes a solution approach to linear problems with many constraints of variable upper bound (VUB) type. This type of constraints are commonly found in various scheduling type problems for which tighter bounds are essential to achieve an efficiency in enumeration. An analytical framework based on factorization is adopted to devise a solution approach to the problem and extend it for more generalized VUB problem (GVUB). This research shows why the VUB type constraints are amenable to the factorization and gives a unified approach to generalized upper bound(GUB) problems, VUB problems and GVUB problems. Implementation issues are also included.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.126-126
• /
• 2000

In this paper, we provide an approach of controller synthesis for input-saturated linear systems by a linear parameter varying (LPV) framework. Using directly the saturation nonlinearity as scheduling parameters, we propose an LPV-stabilizer with parameter-dependent dynamic state-feedback controller concept. Especially, the synthesis conditions are formulated in terms of linear matrix inequalities (LMIs) that can be solved very efficiency.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.04a
• /
• pp.119-121
• /
• 2000

실시간 시스템은 설계단계에서 주어진 태스크 집합에 맞게 구성되어 진다. 이러한 상황은 결국 동적인 상황에 시스템을 쉽게 적용할 수 없는 결과를 초래한다. 본 논문에서는 RTLinux환경에서 보다 동적인 상황에 적응하도록 효율적인 태스크 스케쥴링을 위한 프레임웍을 제시한다. Schedulability manager는 실시간 태스크의 특성에 맞게 스케쥴러를 선택하도록 하였으며 Slack manager는 non-critical 태스크를 스케쥴하도록 한다. 이로서 critical 태스크의 마감시한을 준수하면서 non-critical 태스크의 응답시간을 최소화한다.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2010.11a
• /
• pp.461-462
• /
• 2010

• ETRI Journal
• /
• v.36 no.2
• /
• pp.197-205
• /
• 2014

This paper proposes a novel framework for virtual content delivery networks (CDNs) based on cloud computing. The proposed framework aims to provide multimedia content delivery services customized for content providers by sharing virtual machines (VMs) in the Infrastructure-as-a-Service cloud, while fulfilling the service level agreement. Furthermore, it supports elastic virtual CDN services, which enables the capabilities of VMs to be scaled to encompass the dynamically changing resource demand of the aggregated virtual CDN services. For this, we provide the system architecture and relevant operations for the virtual CDNs and evaluate the performance based on a simulation.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2005.05a
• /
• pp.523-527
• /
• 2005

현재 기업에서는 생산 효율을 극대화시키기 위하여 많은 비용을 투자하여 Package형태의 솔루션(S/W)을 도입하고 있다. 그러나 솔루션들은 그 특성상 기업특성을 생산 일정계획 수립에 충분히 반영하지 못하고 있어 실제 도입 후에 사용에 어려움이 겪고 있으며 일부 기업에서는 생산일정 전문가를 통하여 재생산계획을 수립하고 있다. 본 연구는 상용화 되고 있는 솔루션에서 제시된 생산일정을 생산일정 전문가가 회사의 특성을 고려하여 재생산일정을 수립하는 단계에 대한 사용자의 패턴을 추출 후 신경망을 통하여 패턴을 학습하여 재생산일정 수립 시 소요되는 시간을 최소화 시키며 기업의 특성을 반영하는 적응적 일정계획 지원 시뮬레이터의Framework을 제시하고자 한다.

• Industrial Engineering and Management Systems
• /
• v.8 no.3
• /
• pp.155-161
• /
• 2009

We focus on discrete event systems with a structure of parallel processing, synchronization, and no-concurrency. We use max-plus algebra, which is an effective approach for controller design for this type of system, for modeling and formulation. Since a typical feature of this type of system is that the initial schedule is frequently changed due to unpredictable disturbances, we use a simple model and numerical examples to examine the possibility of applying the concepts of the feeding buffer and the project buffer of critical chain project management (CCPM) on max-plus linear discrete event systems in order to control the occurrence of an undesirable state change. The application of a CCPM-based framework on a max-plus linear discrete event system was proven to be effective.