Research Experience
Optimized Coordination and Scheduling of Traffic Evolving on Complex Guide-path Networks, Aug. 2020 - Present.
This research program aims to define new task allocation and robot routing problems in the networked mobile robot systems with the constricted nature of their operational environment and the coordination among robots maintaining full connectivity of a multi-hop wireless communication network.
Provide a systematic and formal description of the considered problems, which (i) culminates in the analytical characterization of these problems in the form of mixed integer programming (MIP) formulations, and (ii) enables the worst-case computational complexity of these problems.
Establish some structural properties of the underlying solution spaces that enable the strengthening of the developed MIP formulations, and introduce strong combinatorial relaxations of the original MIP formulations that (1) for every solution of the proposed relaxations, there always exists a solution for the original MIP formulation with the same objective value, and (2) the optimal solution for the proposed relaxation can be easily converted to optimal traffic schedules for the corresponding problems of the original formulation.
Employ the attained analytical results towards the development of a heuristic algorithm for the considered problems that can provide good-quality solutions for the larger problem instances in a computationally efficient manner.
Development of algorithms for optimal control of a steam power plant, Dec. 2017 - May. 2018.
The purpose of this project is to improve the productivity of power generation by combustors through the development of a smart method to control the operation of combustors in a steam power plant.
Implement statistical analysis to understand relationships among pre-defined features and between features and the amount of generated power and the resulting unit price, and to select a subset of features that are related to the operational effectiveness.
Develop the random-forest model leveraging the analytical results, to enable the smart control of the operation of the burner in each combuster.
The implementation of the developed method improves the operational productivity of a steam power plant. Furthermore, this work is presented in the research competition in POSCO, and awarded second place.
Methodology for dispatching rules' weights, Jul. 2018 - Sep. 2018.
The purpose of this project is to develop a framework for KPI analysis with various weights on pre-defined dispatching rules, and to provide the optimal distribution of weights for each key performance indicator (KPI).
Implement data analysis to understand the relationship between various dispatching rules and the performance in the production system with respect to each KPI.
Extract important features that are related to each KPI, leveraging the analytical results and the interview with the production manager and experts in the production line.
Develop various intelligent models like an artificial neural network model and a random forest model to improve the operational productivity with respect to the pre-defined KPIs, and suggest the methods to control the dynamic weight set for the dispatching rules.
Development of advanced operations management system for smart factory with clean energy, Aug. 2017 - Jul. 2018.
The purpose of this project is to develop a management system for the quality and quantity of the products, through a real-time monitoring and anomaly detection system and an Advanced Planning and Scheduling System (APS).
Provide a systematic introduction of the scheduling problem in the considered manufacturing system in the form of mathematical programming (MP) formulation to evaluate the performance of the developed APS.
Compare the results between an optimal solution attained by the developed MP formulation and the schedule generated by APS, and modify the planning and scheduling system by adding rules to improve the performance of the generated schedule.
Conduct the statistical analysis to understand the features which may cause the machine disruptions in the production line.
Development of intelligent production management system for smart factory manufacturing parts, Mar. 2017 - Apr. 2018.
This project aims to design a cyber physical system (CPS)-based factory operations monitoring system with the purpose of detecting unexpected events and reacing to the events in an antomotive parts manufacturing factory.
Construct a digital twin factory that enables real-time factory monitoring based on production data from a manufacturing execution system and installed sensors.
Define several unexpected disruptions that affect operational productivity, and derive the criteria with which to automatically detect and diagnose the disruptions.
Develop a rescheduling method considering the defined disruptions to generate a new schedule that increases the production quantity for a given time horizon.
The implementation of the developed monitoring system and schedule modification methods improves the operational productivity with respect to the production quantity.