多階段決策過(guò)程( multistep decision process )是指 這樣一類特殊的活動(dòng)過(guò)程,過(guò)程可以按時(shí)間順序分解成若干個(gè)相互聯(lián)系的階段,在每一個(gè)階段都需要做出決策,全部過(guò)程的決策是一個(gè)決策序列。 動(dòng)態(tài)規(guī)劃 ( dynamic programming )算法 是解決 多階段決策過(guò)程最優(yōu)化問(wèn)題 的一種常用方法,難度比較大,技巧性也很強(qiáng)。利用動(dòng)態(tài)規(guī)劃算法,可以優(yōu)雅而高效地解決很多貪婪算法或分治算法不能解決的問(wèn)題。動(dòng)態(tài)規(guī)劃算法的基本思想是:將待求解的問(wèn)題分解成若干個(gè)相互聯(lián)系的子問(wèn)題,先求解子問(wèn)題,然后從這些子問(wèn)題的解得到原問(wèn)題的解; 對(duì)于重復(fù)出現(xiàn)的子問(wèn)題,只在第一次遇到的時(shí)候?qū)λM(jìn)行求解,并把答案保存起來(lái),讓以后再次遇到時(shí)直接引用答案,不必重新求解 。動(dòng)態(tài)規(guī)劃算法將問(wèn)題的解決方案視為一系列決策的結(jié)果,與貪婪算法不同的是,在貪婪算法中,每采用一次貪婪準(zhǔn)則,便做出一個(gè)不可撤回的決策;而在動(dòng)態(tài)規(guī)劃算法中,還要考察每個(gè)最優(yōu)決策序列中是否包含一個(gè)最優(yōu)決策子序列,即問(wèn)題是否具有最優(yōu)子結(jié)構(gòu)性質(zhì)。
標(biāo)簽: multistep decision process 過(guò)程
上傳時(shí)間: 2015-06-09
上傳用戶:caozhizhi
soft.studa.com大家一起加油啊,好痛苦啊,呵呵
上傳時(shí)間: 2014-06-14
上傳用戶:ukuk
gcc支持soft fp 和hard fp兩種,這里是實(shí)現(xiàn)了soft fp.
上傳時(shí)間: 2014-01-01
上傳用戶:aix008
Turbo Decoder Release 0.3 * Double binary, DVB-RCS code * Soft Output Viterbi Algorithm * MyHDL cycle/bit accurate model * Synthesizable VHDL model
標(biāo)簽: Algorithm Decoder DVB-RCS Release
上傳時(shí)間: 2015-07-10
上傳用戶:清風(fēng)冷雨
machine learning, accuracy estimation, cross-validation, bootstrap, ID3, decision trees, decision graphs, naive-bayes, decision tables, majority, induction algorithms, classifiers, categorizers, general logic diagrams, instance-based algorithms, discretization, lazy learning, bagging, MineSet.
標(biāo)簽: decision cross-validation estimation bootstrap
上傳時(shí)間: 2015-07-26
上傳用戶:趙云興
matlab.soft.base介紹了matlab的軟件基礎(chǔ)和功能使用
上傳時(shí)間: 2013-12-21
上傳用戶:lizhen9880
Find a classification error for a given decision surface D and a given set of patterns and targets
標(biāo)簽: given classification and decision
上傳時(shí)間: 2013-12-18
上傳用戶:xinzhch
support vector classification machine % soft margin % uses "kernel.m" % % xtrain: (Ltrain,N) with Ltrain: number of points N: dimension % ytrain: (Ltrain,1) containing class labels (-1 or +1) % xrun: (Lrun,N) with Lrun: number of points N: dimension % atrain: alpha coefficients (from svcm_train on xtrain and ytrain) % btrain: offest coefficient (from svcm_train on xtrain and ytrain) % % ypred: predicted y (Lrun,1) containing class labels (-1 or +1) % margin: (signed) separation from the separating hyperplane (Lrun,1
標(biāo)簽: classification support machine Ltrain
上傳時(shí)間: 2015-09-04
上傳用戶:問(wèn)題問(wèn)題
請(qǐng)下載方正Apabi Reader進(jìn)行閱讀。http://www.skycn.com/soft/5531.html
標(biāo)簽: Reader Apabi skycn 5531
上傳時(shí)間: 2015-09-05
上傳用戶:czl10052678
Hard-decision decoding scheme Codeword length (n) : 31 symbols. Message length (k) : 19 symbols. Error correction capability (t) : 6 symbols One symbol represents 5 bit. Uses GF(2^5) with primitive polynomial p(x) = X^5 X^2 + 1 Generator polynomial, g(x) = a^15 a^21*X + a^6*X^2 + a^15*X^3 + a^25*X^4 + a^17*X^5 + a^18*X^6 + a^30*X^7 + a^20*X^8 + a^23*X^9 + a^27*X^10 + a^24*X^11 + X^12. Note: a = alpha, primitive element in GF(2^5) and a^i is root of g(x) for i = 19, 20, ..., 30. Uses Verilog description with synthesizable RTL modelling. Consists of 5 main blocks: SC (Syndrome Computation), KES (Key Equation Solver), CSEE (Chien Search and Error Evaluator), Controller and FIFO Register.
標(biāo)簽: symbols length Hard-decision Codeword
上傳時(shí)間: 2014-07-08
上傳用戶:曹云鵬
蟲蟲下載站版權(quán)所有 京ICP備2021023401號(hào)-1