• Journal of the Korea Society of Computer and Information
• /
• v.26 no.12
• /
• pp.29-43
• /
• 2021

In this paper, a new approach to solve the constrained dynamic programming is proposed by using the constraint programming. While the conventional dynamic programming scheme has the state space augmented with states on constraints, this approach, without state augmentation, represents states of constraints as domains in a contraining programming solver. It has a hybrid computational mechanism in its computation by combining solving the Bellman equation in the dynamic programming framework and exploiting the propagation and inference methods of the constraint programming. In order to portray the differences of the two approaches, this paper solves a simple version of the long-term car rental financing problem. In the conventional scheme, data structures for state on constraints are designed, and a simple inference borrowed from the constraint programming is used to the reduction of violation of constraints because no inference risks failure of a solution. In the hybrid approach, the architecture of interface of the dynamic programming solution method and the constraint programming solution method is shown. It finally discusses the advantages of the proposed method with the conventional method.

• Proceedings of the KOR-KST Conference
• /
• 1998.10a
• /
• pp.252-260
• /
• 1998

본 논문은 기개발된 간선도로 연동하 신호최적화 모형인 KS-SIGNAL의 최적화 수행속도를 향상하기 위한 새로운 모형식 및 방법론을 제시하고 평가하였다. 본 논문에서는 탑재, 좌회전 현시순서에 관한 제약식 추가 등의 3단계 작업을 실시하였다. 첫 번째 단계인 모형식의 수정에 있어서는 기존의 모형식에서 변수로 사용하던 대가차량 소거시간을 상수로 산정함으로써 일부 제약식 및 변수를 소거시킬 수 있었으며, 두 번째 단계에서는 선형계획식의 해를 구하기 위한 툴로 사용되는 Wondow용 Lindo library를 탑재, 새로이 변형된 형식의 모형식을 제안한다. 마지막으로 세 번째 단계에서는 좌회전 현시 순서에 관한 제약식을 추가함으로써 최적화 작업에 대한 경우의 수를 줄임으로써 수행속도를 향상시키는 방법론에 대해 제시한다. 결론적으로 기존의 KS-SIGNAL과 비교해 최적화 수행속도는 99%이상 향상되었으며, 도출된 해 또한 타 모형식과 비교해 우수한 결과를 나타냈다.

• Information Systems Review
• /
• v.5 no.2
• /
• pp.203-217
• /
• 2003

Constraint Programming and Optimization have developed in different fields to solve common problems in real world. In particular, constraint propagation and linear Programming are their own fundamental and complementary techniques with the potential for integration to benefit each other. This intersection has evoked the efforts to combine both for a solution method to combinatorial optimization problems. Attempts to combine them have mainly focused on incorporating either technique into the framework of the other with traditional models left intact. This paper argues that integrating both techniques into an old modeling fame loses advantages from another and the integration should be molded in a new framework to be able to exploit advantages from both. The paper propose a declarative modeling framework in which the structure of the constraints indicates how constraint programming and optimization solvers can interact to solve problems.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2007.11a
• /
• pp.468-474
• /
• 2007

가상기업(virtual enterprise)은 21세기에 고객의 요구가 매우 다양해지고 기술은 점점 더 복잡해지는 상황에서 정보네트워크 기술의 발전을 이용한 새로운 기업간 협력을 통해 극복하려는 경쟁전략의 일환이다. 현재까지의 가상기업에서는 주로 EDI(Electronic Data Interchange)를 이용한 국한된 정보의 교류가 이루어졌다. 그러나 이러한 국한된 정보이외에도 실제 가상조직 내에서는 규칙 또는 제약식 형태의 정보의 교류 또한 필요한 실정이다. 이러한 정보의 교류에 시맨틱 웹 플랫폼을 이용하게 되면 더욱 쉽고 유연한 정보의 전달이 가능하다. 현재 웹 온톨로지 언어인 OWL과 시맨틱 웹 규칙언어인 SWRL을 통해서 사실 및 규칙의 포맷은 갖추어져 있는 상황이지만 제약식의 표현에 있어서는 뚜렷한 포맷이 정해져 있지 않아 가상조직 내에서의 구성원간의 제약조건의 교류에 문제가 있는 상황이다. 본 논문에서는 가상기업조직에 있어서 각각의 구성 기업들간의 제약식 공유가 이루어 질 수 있도록, 시맨틱 웹 표현의 대표적 언어인 OWL 기반으로 하는 제약식 표현언어(Semantic Web Constraint Language)를 응용한다. 또한 실제로 이러한 SWCL이 가상기업 문제에 얼마나 효율적으로 적용될 수 있는지 증명하기 위해 실례를 적용해 보고자 한다.

• Journal of Intelligence and Information Systems
• /
• v.10 no.2
• /
• pp.165-187
• /
• 2004

Constraint Programming in AS and Optimization in OR started and have grown in different backgrounds to solve common decision-making problems in real world. This paper tries to integrate results from those different fields by suggesting a hybrid solver as an integration framework. Starting with an integrating modeling language, a way to implement a hybrid solver will be discussed using Warren's abstract machine and an finite domain constraint programming solver structures. This paper will also propose some issues rising when implementing the hybrid solver and provide their solutions.

• Journal of the Korean Data and Information Science Society
• /
• v.6 no.2
• /
• pp.17-27
• /
• 1995

We propose a framework for representing and processing uncertain knowledge on the basis of constraint satisfaction. A system of equations and/or inequalities can be considered as a set of constraints that should be solved, and each constraint in the set is transformed into a corresponding logical formula which can be solved through a constraint solving program. Most of rule-based systems, for instance, use a simple probabilistic theory in order to maintain uncertain knowledge, therefore uncertain knowledge can be represented and processed in the constraint satisfaction program quite efficiently.

• Journal of Korea Spatial Information System Society
• /
• v.11 no.1
• /
• pp.79-86
• /
• 2009

The tuple of constraint database consists of constraint logical formula and it could process the presentation and query of the constraint database simply. Query operation processing shall include the constraint formula between related tuple such as selection, union, intersection of spatial data through the constraint database. However, this could produce the increasing of duplicated or unnecessary data. Hence, it will drive up the cost as per quality. This paper identified problems regarding query processing result in the constraint database. Also this paper suggested the tuple minimization summary method for result relation and analyzed the effects for efficient query processing. We were able to identify that the effectiveness of the query processing was enhanced by eliminating unnecessary constraint formula of constraint relation using the tuple minimization method.

• Journal of the Korea Convergence Society
• /
• v.10 no.11
• /
• pp.71-77
• /
• 2019

This paper defines a multi-selection knapsack problem and presents an algorithm for seeking its optimal solution. Multi-selection means that all members of the particular group be selected or excluded. Our branch-and-bound algorithm introduces a simplex containing the feasible region of the original problem to exploit the fact that the most tightly underestimating function on the simplex is linear. In bounding operation, the subproblem defined over the candidate simplex is minimized. During the branching process the candidate simplex is splitted into two one-less dimensional subsimplices by being projected onto two hyperplanes. The approach of this paper can be applied to solving the global minimization problems under various types of the knapsack constraints.

• The Journal of Engineering Geology
• /
• v.28 no.3
• /
• pp.367-377
• /
• 2018

This study introduces a DC resistivity inversion method that incorporates structural and inequality constraints to enhance the resolution of resistivity inversions, and presents sample inversion results with these constraints. In the constrained inversions, a base model is constructed from a layered model through interpretation of other geophysical data. Inversion tests establish that both the structural and inequality constraints produce better resistivity models than the unconstrained inversion. However, the inequality inversion not only reproduces the exact layered structure of the background, it reproduces conductive anomalies at a depth of ~ 10 m when an inexact base model of electrical resistivity is used.

• Journal of Energy Engineering
• /
• v.23 no.4
• /
• pp.150-159
• /
• 2014

Since the release of mid-term domestic GHG goals until 2020, in 2009, some various GHG reduction policies have been proposed. There are two types of modeling approaches for identifying options required to meet greenhouse gas (GHG) abatement targets and assessing their economic impacts: top-down and bottom-up models. Examples of the bottom-up optimization models include MARKAL, MESSAGE, LEAP, and AIM, all of which are developed based on linear programming (LP) with a few differences in user interface and database utilization. In this paper, we suggest a simplified LP formulation and how can build it through step-by-step procedures.