【正文】
at least isolate the parts of your AI that are unpredictable and more difficult to develop ,test,and debug,while ideally keeping the majority of your AI system in tranditional form.這些成功的游戲使用非定性方法與更傳統(tǒng)的定性方法結(jié)合起來(lái),并將這些方法只用在需要的地方,并且與唯一問題形成最佳匹配。9。這一方法我們提倡使用非定性方法。這些游戲的成功引發(fā)了新的學(xué)習(xí)AI方法的興趣,如決策樹算法、神經(jīng)網(wǎng)絡(luò)算法、遺傳算法和概率的方法。此外,它讓人更難以真正理解AI究竟在做什么,這使得調(diào)試更加困難。s more difficult to really understand what the AI is doing, which makes debugging more difficult. These factors have proven to be serious barriers for widespread use of learning AI techniques.在游戲AI中,下一件大事就是學(xué)習(xí)?;旧?,一個(gè)生命系統(tǒng)是一個(gè)綜合系統(tǒng),具有自然的行為。在一個(gè)第一人稱射擊類游戲中,被射擊者需要穿過地牢或建筑以完成接近或逃避玩家控制的人物的行為。Effective and efficient pathfinding is a fundamental task that nonplayer characters must acplish in all sorts of games. Nonplayer character units in a war simulation must be able to navingate over terrain an aviod barriers to reach the in a firstperson shooter must be able to navigate through dungeons or buildings to reach or escape from the player. The scenarios are endless, and it39。t the only wellestablished AI technique. Finite state machines are a ubiquitous game AI technique, but basically the idea is to enumerate a bunch of actions or states for putercontrolled characters and execute them or transition between them using ifthen conditionals that check various conditions and criteria.當(dāng)然,欺騙不是唯一行之有效的AI技術(shù)。此外,不平衡的欺騙將使電腦過于強(qiáng)大,使玩家無(wú)法擊敗電腦。例如,在一個(gè)戰(zhàn)爭(zhēng)策略型游戲中,電腦控制的軍隊(duì)不用像游戲人必須做的那樣放出偵察兵,就能夠知道其對(duì)手也就是玩家操作的軍隊(duì)的位置、數(shù)量兵種等一系列信息。事實(shí)證明,這種集中于更好的和更快的圖形技術(shù),包括硬件加速,為開發(fā)更好更復(fù)雜的AI(軟件)現(xiàn)在會(huì)需要更多的資源。t do something silly in some case? Game developers face an evershortening development cycle that makes developing and testing new technology to productionready standards extremely difficult. Such short developlment periods make it difficult for develpers to understand cuttingedge AI technologies fully and to see their implications in a massmarket mercial game.傳統(tǒng)的開發(fā)人員一直對(duì)采用非定性AI持謹(jǐn)慎的態(tài)度,盡管這些情況正在改變。此外,游戲開發(fā)者不用把所有可能的情況全部用明確的編碼去實(shí)現(xiàn)。此外,定性方法不利于學(xué)習(xí)和發(fā)展,并且玩家在玩過一段時(shí)間后很容易找到游戲的規(guī)律,從而影響游戲的可玩性。 shoulders .Further,deterministic methods do not facilitate learning or evolving. And after a litter gameplay, deterministic behaviors tend to bee predictable. This limits a game39。這些行為具有一定程度的不確定性并且是不可預(yù)測(cè)性(不確定性的程度取決于AI方法的應(yīng)用以及這些方法的執(zhí)行情況)。不含有任何的不確定性。Deterministic Versus Nondeterministic AI定性與非定性的AIGame AI techniques generally e in two flavors: deterministic and nondeterministic.游戲AI技術(shù)一般有兩種形式:定性和不定性。雖然游戲AI經(jīng)常被用來(lái)解決相當(dāng)復(fù)雜的問題,但我們可以利用AI去嘗試給機(jī)器方以個(gè)性模樣、不同的性格,或塑造情感和各種性情例如,害怕,焦慮,等等。In games, we aren39。re going to stick with a broad interpretation of game AI, which includes everything from simple chasing and evading, to pattern movenment, to neural networks and genetic algorithms. Game AI probably best fits within the scope of weak AI, However, in a sense you can think of game AI in even broader terms. 一些開發(fā)商認(rèn)為尋路算法是游戲AI的一部分,有些甚至認(rèn)為碰撞檢測(cè)是游戲AI的一部分。例如,開發(fā)人員多年來(lái)都使用AI來(lái)給游戲中無(wú)數(shù)的人物以生命力,從經(jīng)典的街機(jī)游戲Pac Man的鬼魂到第一人稱射擊游戲中的機(jī)器人,還有很多其他游戲。在一些結(jié)構(gòu)比較復(fù)雜,句子比較長(zhǎng)的文章中,只要我們細(xì)心分析句子結(jié)構(gòu),分解句子成分,分清主謂賓,以及哪些部分修飾哪個(gè)部分,我們便可以比較輕松地理解句子內(nèi)容,并將其翻譯過來(lái),當(dāng)然有足夠的詞匯量是必須的。所以,專業(yè)英語(yǔ)的學(xué)習(xí)是很必要的。 雖然我們已經(jīng)學(xué)習(xí)了很長(zhǎng)一段時(shí)間的英語(yǔ),但還遠(yuǎn)遠(yuǎn)不能達(dá)不到輕松地翻譯一篇專業(yè)英語(yǔ)的水平。先說(shuō)明什么是定性AI和非定性AI,然后敘述定性AI和非定性AI在游戲中的應(yīng)用,并且重點(diǎn)說(shuō)明非定性AI的應(yīng)用。這是我提倡的使用這些非定性方法的用法。這些游戲的成功引發(fā)了人們學(xué)習(xí)AI方法的新熱流,如對(duì)決策樹、神經(jīng)網(wǎng)絡(luò)、遺傳算以及概率統(tǒng)計(jì)的方法產(chǎn)生興趣。此外,它讓人更難以真正理解AI究竟在做什么,這使得調(diào)試更加困難。游戲AI的展望游戲