Sung-Seok Ko

Simulation-based two-phase genetic algorithm for the capacitated re-entrant line scheduling problem

In this paper, we develop a simulation-based two-phase genetic algorithm for the capacitated re-entrant line scheduling problem. The structure of a chromosome consists of two sub-chromosomes for buffer allocation and server allocation, respectively, while considering all possible states of the system in terms of buffer levels of workstations and assigning a preferred job stage to each component of the chromosome. As an implementation of the suggested algorithm, a job priority-based randomized policy is defined, which reflects the job priority and the properness of local non-idling in allocating buffering and processing capacity to available job instances. The algorithm is combined with a polynomial time sub-optimal deadlock avoidance policy, namely, Bankers algorithm, and the fitness of a chromosome was evaluated based on simulation. The performance of the proposed algorithm is evaluated through a numerical experiment, showing that the suggested approach holds considerable promise for providing effective and computationally efficient approximations to the optimal scheduling policy that consistently outperforms the typically employed heuristics.

Analyzing technology impact networks for R&D planning using patents: combined application of network approaches

Recently, national governments have tried to improve technology ecology, by formulating research and development (R&D) policies and investing in R&D programs. For strategically designed national R&D plans, analytic approaches that identify and assess the impact of each technology from short-term and long-term perspectives are necessary. Further, in methodological perspective, the approaches should be able to synthetically consider the most recent technological information, the direct and hidden impacts among technologies, and the relative impacts of the focal technology in globally-linked technological relationship from the overall perspective. However, most previous studies based patent citation networks are insufficient for these requirements. As a remedy, we present a combined approach for constructing a technology impact network and identifying the impact and intermediating capability of technology areas from the perspective of a national technology system. To construct and analyze the technology impact network, our method integrates three network techniques: patent co-classification (PCA), decision making trial and evaluation laboratory (DEMATEL), and social network analysis (SNA). The advantages of the proposed method are threefold. First, it identifies the directed technological knowledge flows from the most recent patents, by employing PCA. Second, the proposed network contains both the direct and indirect impacts among different technology areas, by applying the DEMATEL method. Third, using SNA, the method can analyze the characteristics of the technologies in terms of the comprehensive impacts and the potential brokerage capabilities. The method is illustrated using all of the recent Korean patents (58,279) in the United States patent database from 2008 to 2012. We expect that our method can be used to provide input to decision makers for effective R&D planning.

Approximations of lead-time distributions in an assemble-to-order system under a base-stock policy

In this paper, we consider an assemble-to-order production system in which a product with multiple components is produced under a base-stock policy. The demand follows a Poisson arrival process, and the production time at each facility is exponentially distributed.Our main results are closed-form formulae for approximating the lead-time distributions in equilibrium. The approximation formulae for the lead-time distributions can be obtained with a linear combination of bounds, in which certain constants are obtained from auxiliary properties and simulation experiments. To determine the quality of the approximations, we obtained sample lead-time distributions from many simulations. Kolmogorov-Smirnov tests show that our formulae are very good fits of the sample distributions. The approximations are good over the entire range of traffic intensities and are even more accurate in light or heavy traffic.

Paint batching problem on M-to-1 conveyor systems

An M-to-1 conveyor system consists of multiple upstream conveyors and a single downstream conveyor. In this paper, we investigate the paint batching problem on M-to-1 conveyor systems with the objective of minimizing setup costs. Our research is motivated by a vehicle re-sequencing problem at a major Korean automotive manufacturer. Setup costs are incurred when two consecutive jobs in the downstream conveyor do not share the same feature. Re-sequencing flexibility is limited by the precedence relationship among jobs in the upstream conveyors. First, we develop a mixed integer linear programming model and propose an efficient dynamic programming (DP) algorithm for a 2-to-1 conveyor system. However, because the suggested DP cannot guarantee optimality in general settings, we propose two efficient genetic algorithms (GAs) to find near optimal solutions. Specifically, we design the reordering operation for making offspring to satisfy the precedence condition. We show that the proposed GAs perform prominently with respect to optimality gap and computation time; thus, they are amenable to environments where solutions must be obtained within tight time constraints. HighlightsWe investigate the paint batching problem on M-to-1 conveyor systems with the objective of minimizing setup costs.We develop a mixed integer linear programming model for paint batching problem on M-to-1 conveyor systems.We propose an efficient dynamic programming (DP) algorithm for a 2-to-1 conveyor system.We suggest efficient two genetic algorithms for general M-to-1 conveyor systems.These algorithms are prominent in terms of computation time and optimality gap.

Cycle times in a serial fork-join network

This paper presents formulas for approximating the distribution of the cycle time of a job in a two-stage fork-join network in equilibrium. The key step is characterizing the departure process from the first node. Statistical tests justify that the approximate distribution is a good fit to the actual one. We discuss related approximations for m-stage networks, and present a formula for approximating the mean cycle time in a m-stage fork-join network.

Development of a machining simulation system using the octree algorithm

The overall goal of this thesis is to develop a new algorithm based on the octree model for geometric and mechanistic milling operation at the same time. To achieve a high level of accuracy, fast computation time and less memory consumption, the advanced octree model is suggested. By adopting the supersampling technique of computer graphics, the accuracy can be significantly improved at approximately equal computation time. The proposed algorithm can verify the NC machining process and estimate the material removal volume at the same time.

Arterial Road Incident Detection Based on Time-Moving Average Method in Bluetooth-Based Wireless Vehicle Reidentification System

AbstractIncident detection algorithms, which are an essential part of traffic management systems, have been studied for several decades, but the research focus has primarily been on algorithms for incident detection on freeways and other free-flowing roads. When applied on arterial roads, the achievement of stable performance and scalability are major challenges when developing an effective incident detection algorithm. In this research, the authors propose an incident detection algorithm that utilizes travel time and traffic volume samples generated from a Bluetooth-based wireless vehicle reidentification system that has been implemented on arterial roads. The proposed algorithm is based on a moving average over time, which can recognize sample travel time and traffic volume patterns resulting from incidents. The use of a moving average overcomes limitations resulting from sparse travel time sample data collected. Within the algorithm, a threshold strategy is applied that makes the algorithm easy to impl...

An Arterial Incident Detection Procedure Utilizing Real-Time Vehicle Reidentification Travel Time Data

Our road and highway system has become a source of multiple types of “big data.” One such type of data is travel time data obtained from vehicle reidentification systems, a type that is increasingly available due to the implementation of existing and relatively new technologies such as license-plate recognition and Bluetooth-based wireless vehicle identification. Travel time data obtained in real time from such systems are used to update estimated travel times displayed on variable message signs, and research has also been conducted that utilizes travel time data as inputs to incident detection algorithms. Implementation of such systems and prior research has primarily focused on freeways and other free-flowing roads. However, such systems for travel time data collection are also being implemented on arterials. In this research an incident detection procedure that utilizes point-to-point travel time data obtained from an arterial vehicle reidentification system is developed and evaluated. Historical travel time data provided by a Bluetooth-based travel time data collection system and reported incident data are utilized to evaluate the procedure. The results show that the procedure provides a good balance of detection and false alarm rates.