• Title/Summary/Keyword: I-Q

Search Result 1,728, Processing Time 0.032 seconds

Reinforcement learning Speedup method using Q-value Initialization (Q-value Initialization을 이용한 Reinforcement Learning Speedup Method)

  • 최정환
    • Proceedings of the IEEK Conference
    • /
    • 2001.06c
    • /
    • pp.13-16
    • /
    • 2001
  • In reinforcement teaming, Q-learning converges quite slowly to a good policy. Its because searching for the goal state takes very long time in a large stochastic domain. So I propose the speedup method using the Q-value initialization for model-free reinforcement learning. In the speedup method, it learns a naive model of a domain and makes boundaries around the goal state. By using these boundaries, it assigns the initial Q-values to the state-action pairs and does Q-learning with the initial Q-values. The initial Q-values guide the agent to the goal state in the early states of learning, so that Q-teaming updates Q-values efficiently. Therefore it saves exploration time to search for the goal state and has better performance than Q-learning. 1 present Speedup Q-learning algorithm to implement the speedup method. This algorithm is evaluated. in a grid-world domain and compared to Q-teaming.

  • PDF

Polynomials satisfying f(x-a)f(x)+c over finite fields

  • Park, Hong-Goo
    • Bulletin of the Korean Mathematical Society
    • /
    • v.29 no.2
    • /
    • pp.277-283
    • /
    • 1992
  • Let GF(q) be a finite field with q elements where q=p$^{n}$ for a prime number p and a positive integer n. Consider an arbitrary function .phi. from GF(q) into GF(q). By using the Largrange's Interpolation formula for the given function .phi., .phi. can be represented by a polynomial which is congruent (mod x$^{q}$ -x) to a unique polynomial over GF(q) with the degree < q. In [3], Wells characterized all polynomial over a finite field which commute with translations. Mullen [2] generalized the characterization to linear polynomials over the finite fields, i.e., he characterized all polynomials f(x) over GF(q) for which deg(f) < q and f(bx+a)=b.f(x) + a for fixed elements a and b of GF(q) with a.neq.0. From those papers, a natural question (though difficult to answer to ask is: what are the explicit form of f(x) with zero terms\ulcorner In this paper we obtain the exact form (together with zero terms) of a polynomial f(x) over GF(q) for which satisfies deg(f) < p$^{2}$ and (1) f(x+a)=f(x)+c for the fixed nonzero elements a and c in GF(q).

  • PDF

A NOTE ON CYCLOTOMIC UNITS IN FUNCTION FIELDS

  • Jung, Hwanyup
    • Journal of the Chungcheong Mathematical Society
    • /
    • v.20 no.4
    • /
    • pp.433-438
    • /
    • 2007
  • Let $\mathbb{A}=\mathbb{F}_q[T]$ and $k=\mathbb{F}_q(T)$. Assume q is odd, and fix a prime divisor ${\ell}$ of q - 1. Let P be a monic irreducible polynomial in A whose degree d is divisible by ${\ell}$. In this paper we define a subgroup $\tilde{C}_F$ of $\mathcal{O}^*_F$ which is generated by $\mathbb{F}^*_q$ and $\{{\eta}^{{\tau}^i}:0{\leq}i{\leq}{\ell}-1\}$ in $F=k(\sqrt[{\ell}]{P})$ and calculate the unit-index $[\mathcal{O}^*_F:\tilde{C}_F]={\ell}^{\ell-2}h(\mathcal{O}_F)$. This is a generalization of [3, Theorem 16.15].

  • PDF

Plants Community Structure Analysis of Lindera erythrocarpa Native Forest in the Central Korea(I) (중부지방 비목나무 자생림의 식물군집구조 분석(I))

  • 이동철;심경구;최송현;이경재
    • Journal of the Korean Institute of Landscape Architecture
    • /
    • v.22 no.2
    • /
    • pp.133-157
    • /
    • 1994
  • This study was executed to find out the succession stage and the ecological niche of Lindera erythrocarpa Markino. Four sites were selected by field investigation. They are Jeondungsa and Jeongsusa of Kanghwa Island, Mt. Suri of Anyang and Mt. Gaya of Chungcheongnamdo. They located in the region which have the similar temperature with Seoul region or lower average temperature for winter than that of adjacent Seoul. In the four sites, L, erythrocarpa was appeard in canopy layer at L. erythrocarpa community in Jeondungsa, L. erythrocarpa-Q. serrata, Z. serrata-L. erythrocarpa community in Jeongsusa, Castanea crenata-L, erythrocarpa community, L. erythrocarpa-Q. serrata community in Mt. Gaya and in the rest of the sites, it lives in subtree and shrub layer. And in the four sites but Jeongsusa area, it correspond with Chang(1991)'s study that L. erythrocarpa is dominant species in the site impacted by human. L. erythrocarpa lives with Quercus spp. such as Q. serrata, Q. variabilis, Q. mongolica and Carpinus laxiflora but it's presumably a passing phenomena.

  • PDF

Differential Quadrature Analysis for Vibration of Wide-Flange Curved Beams (D.Q.M.을 이용한 I-단면 곡선보의 진동해석)

  • Ji-Won Han;Ki-Jun Kang
    • Journal of the Korean Society of Safety
    • /
    • v.13 no.3
    • /
    • pp.163-170
    • /
    • 1998
  • The differential quadrature method (D.Q.M.) is applied to computation of eigenvalues of small-amplitude free vibration for horizontally curved beams including a warping contribution. Fundamental frequencies are calculated for a single-span, curved, wide-flange beam with both ends simply supported or clamped, or simply supported-clamped end conditions. The results are compared with existing exact solutions and numerical solutions by other methods for cases in which they are available. The differential quadrature method gives good accuracy even when only a limited number of grid points is used.

  • PDF

A Study on the Ultimate Point Resistance of Rock Socketed Drilled Shafts Using FLAC3D and UDEC (유한차분해석과 개별요소해석을 이용한 암반에 근입된 현장타설말뚝의 선단지지력 연구)

  • Lee, Jae-Hwan;Cho, Hoo-Yeon;You, Kwang-Ho;Jeong, Sang-Seom
    • Journal of the Korean Geotechnical Society
    • /
    • v.28 no.1
    • /
    • pp.29-39
    • /
    • 2012
  • The maximum unit point resistance ($q_{max}$) of rock socketed drilled shafts subjected to axial loads was investigated by a numerical analysis. A 3D Finite Difference Method (FDM) analysis and a Distinct Element Method (DEM) analysis were performed with varying rock elastic modulus (E), discontinuity spacing ($S_j$), discontinuity dip angle ($i_j$), and pile diameter (D). Based on the results of obtained, it was found that the ultimate point resistance ($q_{max}$) increased as rock elastic modulus (E) and rock discontinuity spacing ($S_j$) increased. But, it was found that $q_{max}$ decreased as pile diameter (D) increased. As for the influence of the dip angle of rock discontinuity ($i_j$), it was shown that $q_{max}$ decreased up to 50% of maximum value within the range of $0^{\circ}$ < $i_j$ < $60^{\circ}$ due to the shear failure at rock discontinuities. Furthermore, it was found that if $20^{\circ}{\leq}i_j{\leq}40^{\circ}$, influence of $i_j$ should be taken into account because $q_{max}$ tended to approach a minimum value as $i_j$ approached a value near the friction angle of the discontinuity (${\phi}_j$).