정승환 교수

Motivated by the recent reshoring initiatives, we study firms’ global sourcing strategies from two different supply sources, offshore and onshore supply bases, under supply and demand uncertainty. Interestingly, even though onshore sourcing has a significant benefit from market responsiveness, offshore sourcing having no cost advantage also provides benefits in the presence of yield uncertainty and market competition. If competing firms commit single sourcing, one firm may choose inferior offsho

Problem definition: We consider opportunities for cooperation at the supply level between two firms that are rivals in the end-product market. One of our firms is vertically integrated (VI), has in-house production capabilities, and may also supply its rival. The other is a downstream outsourcing (DO) firm that has better market information. The DO is willing to consider a supply partnership with the VI, but it also has the option to use the outside supply market. Academic/practical relevance: S

In this paper, we introduce and study the joint inventory selection and online resource allocation problem, which is characterized by two sequential sets of decisions that are irrevocably linked. First, a decision maker (DM) must select starting inventory levels for a set of available resources. Subsequently, the DM must match arriving customers to available resources in an online fashion so as to maximize expected reward. We first study the problem in its most general form, before focusing on a

Paper presents a Ku-band FMCW radar transceiver IC realized in 0.13 pm CMOS processes. In the radar receiver, a sensitivity time control using a DC offset cancellation feedback loop is employed, which preserves the receiver's SNR not depending on the distance. The radar receiver achieves the full chain gain of 82 dB, P1dB of -2.0 dBm at the minimum gain, and noise figure of 7.9 dB with 106 dB dynamic range. The measured result of the radar transmitter reveals 9 dBm output power. The radar transc

최근 알고리즘 트레이딩에 대한 관심이 높아지면서, 인공지능 방법론을 이용한 매매 전략 구축에관련된 연구들이 활발하게 진행되고 있다. 하지만 복수의 인공지능 방법론을 융합하여 매매 전략 개발에 이용한 사례는 아직 많지 않다. 본 연구는 주가지수선물시장을 바탕으로 인공지능 방법론 중 하나인 러프집합 이론을 적용하여 알고리즘 트레이딩 매매전략을 개발한다. 특히 유전자 알고리즘을 도입하여 생성된 매매전략을 현재시장상황에 최고의 수익률을 보일 수 있도록 최적화한다. 실증분석으로는 2009년부터 2012년까지 4년간의 매매수익률을 분석한 결과 매수 후 보유 전략과 비교하여 우수한 성과를 보였다.

A radar sensor for high speed moving objects is implemented using single chip CMOS UWB IC. To reduce the measuring and processing time for the fast moving target, the reliable and low power consumed envelope detection architecture is used at the expense of high resolution. The equivalent time sampling technique is adopted for UWB transceiver, employing delay locked loop and digitally synthesized impulse generator with 1 ns resolution to detect the fast moving target. The measured results with 4

A 3 - 5 GHz UWB radar chip for surveillance and reconnaissance applications is fabricated in 0.13um CMOS process. The UWB radar chip employs the equivalent time sampling and 4-channel time interleaved sampler architecture. The DC offset calibrator and BJT based circuit greatly reduce the DC offset issues. The analog signal processing part employs the proposed high speed track and hold (T/H) circuit. The direct loop back test shows the feasibility of the distance sensing with 3 cm resolution at t

We study the impact of government financing designed to promote clean technology advancement. Government financing offers loans for financially constrained firms that manufacture and market clean-technology products. We build an analytical model to explore the impact of such government financing in the presence of market uncertainty. Our analysis shows that compared to prevalent commercial financing schemes such as bank financing and equity financing, government financing encourages firms to pur

A fully integrated impulse based UWB receiver system is designed and simulated using 0.18um CMOS technology. The UWB receiver is the non-coherent scheme that simplifies the RF architecture and reduces power consumption. The receiver consists of 3-stage LNA, envelop detector, VGA, comparator and pulse shaper. RF front-end of the system achieves noise figure under 3dB from 3.1GHz to 5.1GHz, IIP3 of -31dBm, gain of 41dB, and input P1dB of -41dBm, respectively. The simulated sensitivity of full rece

Feature descriptors of point clouds are used in several applications, such as registration and part segmentation of 3D point clouds. Learning discriminative representations of local geometric features is unquestionably the most important task for accurate point cloud analyses. However, it is challenging to develop rotation or scale-invariant descriptors. Most previous studies have either ignored rotations or empirically studied optimal scale parameters, which hinders the applicability of the met