• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.171-171
• /
• 2020

Decision making strategies should consider both adaptiveness and robustness in order to deal with two main characteristics of climate change: non-stationarity and deep uncertainty. Especially, robust strategies are different from traditional optimal strategies in the sense that they are satisfactory over a wider range of uncertainty and may act as a key when confronting climate change. In this study, a new framework named Robust Stochastic Dynamic Programming (R-SDP) is proposed, which couples previously developed robust optimization (RO) into the objective function and constraint of SDP. Two main approaches of RO, feasibility robustness and solution robustness, are considered in the optimization algorithm and consequently, three models to be tested are developed: conventional-SDP (CSDP), R-SDP-Feasibility (RSDP-F), and R-SDP-Solution (RSDP-S). The developed models were used to derive optimal monthly release rules in a single reservoir, and multiple simulations of the derived monthly policy under inflow scenarios with varying mean and standard deviations are undergone. Simulation results were then evaluated with a wide range of evaluation metrics from reliability, resiliency, vulnerability to additional robustness measures. Evaluation results were finally visualized with advanced visualization tools that are used in multi-objective robust decision making (MORDM) framework. As a result, RSDP-F and RSDP-S models yielded more risk averse, or conservative, results than the CSDP model, and a trade-off relationship between traditional and robustness metrics was discovered.

• Journal of the military operations research society of Korea
• /
• v.11 no.1
• /
• pp.87-110
• /
• 1985

This paper presents optimum policy of water supply distribution of the Osaka Prefecural Waterworks System located in the midwest of Japanese Islands. Owing to the ever increasing demand for water, the Osaka Prefectural Government endeavors to expand potable and industrial water distribution system to satisfy the growing water demand of the constituents under its jurisdiction. In this regard, the paper discusses a problem of establishing an efficient and effective water distribution system. The criteria to be considered are stability of water level at the reservoirs, stability of flow in the network, and the water treatment and distribution cost. These objective functions may be combined to form a multiple objective optimization problem or may be used independently and formulated into single objective optimization problems.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2006.11a
• /
• pp.618-629
• /
• 2006

This Paper is about optimizing preventive maintenance period of connected (r,s) out of(m,n) : F lattice system that one of multi-component system, (m,n) matrix failure of whole system is occurrence when parts that belong in (r,s) matrix part procession of parts arranged with procession are breakdown all. The preventive maintenance about system is very important viewing from system reliability and operational expense viewpoint. Preventive maintenance that misses a time calls big loss by system failure and expense of frequent full equipment is paid excessively in preventive maintenance itself but expense is paid much in preventive maintenance itself and whole expense escalation can be achieved preferably. Through this research, reliability model is constructed that do expense by smallest under full equipment policy chosen through comparison of each full equipment policy and preventive maintenance expense full equipment cycle and r ,s value are made using simulated annealing algorithm and simulated annealing algorithm that converge fast in multi-component system certified most suitable to optimization decision

• Journal of Korea Multimedia Society
• /
• v.21 no.8
• /
• pp.969-981
• /
• 2018

Recent advances in machine learning approaches such as deep neural network and reinforcement learning offer significant performance improvements in generating detailed and varied motions in physically simulated virtual environments. The optimization methods are highly attractive because it allows for less understanding of underlying physics or mechanisms even for high-dimensional subtle control problems. In this paper, we propose an efficient learning method for stochastic policy represented as deep neural networks so that agent can generate various energetic motions adaptively to the changes of tasks and states without losing interactivity and robustness. This strategy could be realized by our novel trajectory search method motivated by the trust region policy optimization method. Our value-based trajectory smoothing technique finds stably learnable trajectories without consulting neural network responses directly. This policy is set as a trust region of the artificial neural network, so that it can learn the desired motion quickly.

• Transportation Technology and Policy
• /
• v.12 no.6
• /
• pp.16-24
• /
• 2015

• Proceedings of the Korean Institute of Communication Sciences Conference
• /
• 1986.10a
• /
• pp.114-118
• /
• 1986

• Proceedings of the Korean Nuclear Society Conference
• /
• 2002.05a
• /
• pp.180.1-180
• /
• 2002

• Journal of Korea Multimedia Society
• /
• v.13 no.12
• /
• pp.1798-1804
• /
• 2010

TCP with AIMD mechanism, one of the most popular protocols in internet, can solve congestion control in wired networks. This protocol, however, is not efficient in wireless networks. This paper proposes a new mechanism namely General AIMD with Congestion Window Upper Bound in which congestion window is limited by an upper bound. By applying optimization theory, we find an optimal policy for congestion window upper bound to maximize network throughput.

• Journal of Korean Institute of Industrial Engineers
• /
• v.34 no.4
• /
• pp.425-432
• /
• 2008

While the Vendor-Managed Inventory(VMI) is a convenient inventory replenishment policy for the customer company, the supplier usually bears the burden of higher inventory and urgent shipments to avoid shortage. Recently some manufacturers begin to fix the production schedule for the next few days (such as three days). Utilizing that information can improve the efficiency of the VMI. In this study, we present a myopic optimization model using a mixed inter programming; and a heuristics algorithm. We compare the performance of the two proposed methods with the existing (s, S) reorder policy. We consider the total cost as the sum of transportation cost and inventory cost at the customer's site. Numerical tests indicate that the two proposed methods significantly reduce the total cost over the (s, S) policy.

• Water Engineering Research
• /
• v.3 no.1
• /
• pp.57-62
• /
• 2002

This research is intended to integrate long-term operation rules and real time operation policy for conservation & flood control in a reservoir. The familiar Yield model has been modified and used to provide long-term rule curves. The model employs linear programming technique under given physical conditions, i.e., total capacity, dead storage, spillways, outlet capacity and their respective elevations to find required and desired minimum storage fur different demands. To investigate the system behavior resulting from the above-mentioned operating policy, i.e., the rule curves, the simulation model was used. Results of the simulation model show that the results of the optimization model are indeed valid. After confirmation of the above mentioned rule curves by the simulation models, gate operation procedure was merged with the long term operation rules to determine the optimum reservoir operating policy. In the gate operation procedure, operating policy in downstream flood plain, i.e., determination of damaging and non-damaging discharges in flood plain, peak floods, which could be routed by reservoir, are determined. Also outflow hydrograph and variations of water surface levels for two known hydrographs are determined. To examine efficiency of the above-mentioned models and their ability in determining the optimum operation policy, Esteghlal reservoir in Iran was analyzed as a case study. A numerical model fur the solution of two-dimensional dam break problems using fractional step method is developed on unstructured grid. The model is based on second-order Weighted Averaged Flux(WAF) scheme with HLLC approximate Riemann solver. To control the nonphysical oscillations associated with second-order accuracy, TVD scheme with SUPERBEE limiter is used. The developed model is verified by comparing the computational solutions with analytic solutions in idealized test cases. Very good agreements have been achieved in the verifications.