多階段決策過程( multistep decision process )是指 這樣一類特殊的活動過程,過程可以按時間順序分解成若干個相互聯(lián)系的階段,在每一個階段都需要做出決策,全部過程的決策是一個決策序列。 動態(tài)規(guī)劃 ( dynamic programming )算法 是解決 多階段決策過程最優(yōu)化問題 的一種常用方法,難度比較大,技巧性也很強。利用動態(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題。動態(tài)規(guī)劃算法的基本思想是:將待求解的問題分解成若干個相互聯(lián)系的子問題,先求解子問題,然后從這些子問題的解得到原問題的解; 對于重復出現(xiàn)的子問題,只在第一次遇到的時候對它進行求解,并把答案保存起來,讓以后再次遇到時直接引用答案,不必重新求解 。動態(tài)規(guī)劃算法將問題的解決方案視為一系列決策的結果,與貪婪算法不同的是,在貪婪算法中,每采用一次貪婪準則,便做出一個不可撤回的決策;而在動態(tài)規(guī)劃算法中,還要考察每個最優(yōu)決策序列中是否包含一個最優(yōu)決策子序列,即問題是否具有最優(yōu)子結構性質。
標簽: multistep decision process 過程
上傳時間: 2015-06-09
上傳用戶:caozhizhi
soft.studa.com大家一起加油啊,好痛苦啊,呵呵
上傳時間: 2014-06-14
上傳用戶:ukuk
gcc支持soft fp 和hard fp兩種,這里是實現(xiàn)了soft fp.
上傳時間: 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
標簽: Algorithm Decoder DVB-RCS Release
上傳時間: 2015-07-10
上傳用戶:清風冷雨
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.
標簽: decision cross-validation estimation bootstrap
上傳時間: 2015-07-26
上傳用戶:趙云興
matlab.soft.base介紹了matlab的軟件基礎和功能使用
上傳時間: 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
標簽: given classification and decision
上傳時間: 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
標簽: classification support machine Ltrain
上傳時間: 2015-09-04
上傳用戶:問題問題
請下載方正Apabi Reader進行閱讀。http://www.skycn.com/soft/5531.html
上傳時間: 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.
標簽: symbols length Hard-decision Codeword
上傳時間: 2014-07-08
上傳用戶:曹云鵬