Showing posts with label SIMC. Show all posts
Showing posts with label SIMC. Show all posts

Sunday, 20 February 2011

偽.SIMC回憶錄5

12:00 midnight, Day 5
不管有沒有慶功的意味,那晚的確玩得痛快。
一副撲克、一場足總盃還有一堆零食,我們足足玩了一整晚(11PM ~ 9AM),從鋤Dee、German、傷心小棧跟Contract Bridge都是取樂的部分之一……事實證明了數學人的確很喜歡這類思考性強的遊戲。
翌日早晨,我們坐那載我們來的旅遊巴回去機場。跟那些當地學生道別實在不容易,就正如這五日已足夠讓我們跟別隊留下深刻友誼一樣。他們道了我們一點小手信(仍掛在床頭)和道別卡,希望這些回憶不是僅僅留在紙上而真是長存心中吧。

飛機起飛,道別了星洲,也宣告著這活動的結束。

Day After I
回學校做演說,沒有甚麼好提的。我只想令自己記得用話筒的正確用法。

Day After II
今天再次拿起合照,腦海中先補上了一句see you in SIMC 2012,再想起各種有趣的回憶,百般滋味非筆墨所能形容。我們隊伍間仍然以fb, msn互相聯繫著,但願這些已能長久地保持下去吧。
回到正題,關於我一開始問自己的問題,我給出的答案是否定的。廁所的話我回程的時候自己親眼看到了被大陸人弄壞的公廁,但市內的衛生一直都不錯,至少街邊的伯伯才不會做出很多不衛生的行為。那邊在家長式鐵腕政策下這方面的確做得不錯。
至於自己的解答方面,我肯定我方的解答是唯一用直觀法做的。但不得不承認一點,Greedy Alogarithm的確是更有效的解答。在嚴謹性方面,我認為我們可以用一些比較容易的方法證明"選取非greedy alogarithm的地段使答案非最大化",不過我還沒有作出任何嘗試。 orz

Extending question
1) Prove / disprove "greedy alogarithm maximizes answer. If not, please give a map that greedy alogarithm does NOT work (and give the optimal answer at the same time).
2) Account the existance for general answer for a map that all street lies on grid square while all intersection points are in integal points.
3) Account the same thing but the map is not necessraily on the grid. Try to solve this in the light of topology concept?
4) Solve these things too, for i) constrains of Q2, ii) constrains where population density is considered.


fin.

完成了。留意上面的圖,3"E"是SIMC的標語,代表Expeior, Expono, Excedo, 也就是Experiment, Explore和Excel,這的確成了我的座右銘,因為它很簡單明確地道出了科學的精神……

*重申,此文純屬虛構,如有雷同實屬巧合。

Wednesday, 2 February 2011

偽.SIMC回憶錄 4.2

12:00 noon, Day 4, NUS High School
口頭報告的完成代表著比賽的時間已完結,剩下來的時間就可以等賽果和輕鬆一下。
今天的午飯不知到是否因為壓力頓消而胃口大開,今餐似乎異常豐富(意粉/蛋餅/布丁?),飲品還是那被再度稀釋的雜果賓治,還有超過飽和份量而好像漿糊的熱朱古力(這令我想起明將XD),兩者之下我還是選擇了飲自己的清水……
由於結果在傍晚才會公佈,中間的時間我們同樣會出外觀光,今次我們去了星洲的Science Park (有點似香港的科學園,但那邊的Science Park真的是一個Park)和濾水廠(這裡有兩個小插曲,第一個:某房間的牆壁有一個圈圈九的圖樣,其內容大概是濾水的第九個步驟,但……相信東方迷會知道另一個意思。
另一個插曲是一某個模仿水壩的模型,其中有一部分是關於下雨儲水,模型由電腦控制,所以會自己來;但當地導游似乎算準了時機(大概他也練過很多次?),大呼一聲Rain!,雨就下來了。
最後一個地方是可持展續發展的展館,最深刻的要算上用數萬個膠樽砌成的球體,以及貌似jubeat,會感應的地磚……
一眨眼時間又到了傍晚時分,我們回到了演講廳等候結果公佈(簡單來說就是看看有沒有獎)。看來星洲政府頗為重視這比賽,頒獎嘉賓為當地教育局局長(雖然這樣說有點不敬,但我隊一致認為其裝扮酷似推點心車的阿姐……XD)。
Raymond Chan的致詞中提到解答技巧的多樣性,例如loop programming, linear programming, linear algebra, matrix, algorithms,,,當然沒有提我們這種直觀解法,於是當時我似乎對自己沒啥信心可以拿到獎,始於放不下自己只是到這裡吸收經驗的事實。
當然,我是在場其中一個笑到最後的人,連同友隊的distinction,我校應該是除了主場的NUS High School以外成績最好的學校(因為他們拿到冠軍),而北京某校(上屆冠軍)啥都拿不到。
頒獎完成後來到活動的終曲:Farewell Party。
五天,不太短亦不太長的時間,能令一個完全不想講英文的人愉快地和各國好手交流,能撮合一個澳洲佬和一個日本妹(?!?),能令各人的數學技巧更上一層樓……最重要的是,SIME為我們留下了不可磨滅的美好回憶。
這一點,從我們不停向四周的對手在合照上簽名可以得證。每個人互相交換MSN,留下對對手的Comment(可惜當時未流行"Like"……),一個個簽名,代表著五日以來建立的友情……如果十年後、二十年後,當我們拿出這張照片時,腦海中還能閃過一張張當年的回憶,那我就很滿足了。
晚會基本上以簽名開始,也以簽名結束。一些隊伍因為趕夜機,已經不見人影。天下無不散之筵席,當大家各自離去,就是活動完滿結束之時。
……

Monday, 10 January 2011

偽.SIMC回憶錄4.1

6:45a.m. , Day 4, NUS High School
相信每一個人都在準備報告而在作出最後努力吧。每隊都如同前幾日,在早餐時段(今天是三絲炒米)仍然打開電腦--當然不是打LF2,而是為簡報作最後衝刺。
上午八時,所有人進入了演講廳,找了幾個位置坐下準備,然後工作人員說明規則:上午一共有七節時間(每節三十五分鐘),當中有四節時間我們將要前往獨立的課室向一位評審作報告。其間簡報內容不得修改(簡報也事先上繳了),以免不同評審所接收的有差別。
第一個評審應該是中國人,我們去的課室從桌子椅子到擺設都很像,這給予我們一個比較舒適的環境,心情也稍為平靜下來。但很快,我們便體會到口述報告的難處:指示上說每次報告約為十五分鐘,加上問答環節才最多二十五分鐘,但很多時候直觀證明會比一種已知的演算法(例如今次大派用場的貪心演算法),在一個相對嚴謹的態度下每個規則都要加以證明,使時間嚴重不足。幸好,超時並不會扣分,加上評審的問題集中在第一二題,我們也自然地將重點放在頭兩題上。第三題的自由發揮上並不用太多證明,於是第三題似乎只用了兩分鐘時間……(後來,原來第三題是我們失分重點 orz)
第二個評審比較特別,是某位本地中學的校長,當地學生也沒有怠慢,千叮萬囑叫我們進去打招呼……果然有地位的人有特別特遇,今次我們的報告地點是電腦實驗室 (ICT Lab),注意這比普通電腦室高級,香港也沒多少學校有這等設施。Lab的後方坐著一位白髮蒼蒼的老人,大概就是我們的評審了吧。當然我們不會因為評審的不同在報告內容方面有所不同,但也變得更緊張了。更要命的是,白板在上一隊用完後有未刷過的痕跡,整個白板都用矩陣語言寫成,這倒給了我們不少壓力。當然,報告還算順利完成(事後我發現Lab太大,我畫的圖根本看不到)……
第三個評審同樣是個本地志授,我們這次的報告並沒有很待別的地方。然後我們有兩節休息的時間,可以回到演講廳作最後準備。我發現了我身後的地板有個隱藏的插頭,於是就拿了來充電,但其實我那部電腦並不是用來做簡報的那部電腦,當時我也沒有意識到這問題的嚴重性,後來卻差點帶來不可挽回的後果……
我們來到第四位評審的門外,看見裡面有一隊在報告(我第一個想法是:超時扣分?),再望了白板一眼,屆然是同樣的矩陣語言!那無疑是經過第二位評審的那一隊,那是來自北京市的某隊,似乎沒有學服,於是全西裝出戰;這是個頗有趣的地方,有不少學校根本沒有學服(以歐洲為甚),今天的指示要求學生穿校服,但沒有學校的隊伍服飾便各有不同:印尼隊選擇穿上民族服飾、英國隊便衣出戰、澳洲隊更只穿上球衣上陣……對愛面子的中國隊來說西裝應該是個不錯的選擇……嗯,下次試旗袍和中山裝?!?
同時,我們也注意到門外有一張SNMO (Singapore National MathO)的海報,這對我們這些奧數玩家來說有不少吸引力,事後我們也在網上找到那年的題目,其實都不難……
言歸正傳,我們進入第四位評審的課室裡報告,我們有題用直觀選擇法做了個比optimal solution少0.5的答案,然後我卻不心口快說了句"This is the best result obtained by our selection moethods.",評審卻可能對上一隊有答案有印象,反問一句"Are you sure this is the best solution",我心知大事不妙也只能硬著頭皮回答"This is the best result obtained by our trials under the selection methods"……
最後一次報告過程頗為驚險,先是投影器的線插不進,然後被挑Starbuck的logo不能隨便放上簡報(一笑置之?),做完簡報那一刻剛好沒電(我當時立刻後悔自己應該讓做簡報那部電腦充電……)
就這樣,四次報告總算有驚無險地完成了。

Friday, 31 December 2010

偽.SIMC回憶錄3.2

6:45, Day 3, NUS high school
今天是繳交written report的日子,我們要做的事就是把昨晚做出來的內容文字化交出去而已。有電腦,有MS word,我們以為要打一份報告十分容易,其實不然。
一台手提電腦,缺掉滑鼠,其機動性就好像減了一半。特別是針對給定的地圖上畫畫作出解釋,這對我們難度十分高,要用手提電腦的觸控版實在太難了。
不過,要說到港人還有甚麼優點的話,那就是夠快了。答案是在昨晚figure out,至於要怎樣下筆,則從來沒有想過。所謂的報告應該有格式無格式,還是好像公開試那樣寫上題號就能作答?我也不知道。但真正坐在電腦前方,身邊的一切就在這一瞬間凝固掉,身處一間沒冷氣室溫三十五度的房間也好,隔壁的中國隊無聊得在跳舞也好,我們唯一看到就是眼前的文件。 (這倒有點像HKMO, 聖賢之類的團體賽,能幾個人圍在一起計一題數的情況真少見)
於是,很快地,這份報告就像一份第一堂收的功課一樣被趕起了。沒有多餘的修飾,只有簡陋的說明圖;沒有華麗的算式,只有簡單直接的解釋。
……
正午是報告的截數時間,從十時開始趕工,到最後我們早了三十分鐘交出來,可說是不錯的速度了。不過,書面報告之後,還有一項更剌激的,就是口述報告。
在第四日早上,我們會隨機被分到四位評審,我們要分四次單獨的解釋我們的報告。在這種問題下,排除窮舉外幾乎沒有「絕對正確」的答案,因為這部分最刺激的地方在於評審有權隨時打斷你的發言來質問你(原則上,Q&Asection是最後五分鐘,但事實上顯然不是)。
此時我們開始預備簡報,問題是,我們的解答中,證明了幾個直觀排列「是開店的必然法則」來證明我們的答案。很可惜,當地圖越來越複雜時,一間店身處於某座標時不能同時滿足所有的必然法則,在「開店開小化」下我們顯然不能滿足所有法則,那在直觀排列下的可能答案就不止一個了。窮舉顯然不是很好的辦法。

晚上我們暫停了手上的工作,選擇了跟各國選手比試一下--當然不是數學,我們選擇了另一種全場通用的語言--體育作為切磋的部分。雖說是比賽途中,但每個人就輕鬆得像在學校跟隊打一樣,我們就這樣渡過了一晚。
順帶一提,這天晚上的食物是類似炒年糕的東西,再配以炸魷魚(煎炸炒,很地道的煮法?),但最神奇的是前幾天桃紅色的賓治被稀釋成淺紅色,據說有人服用後腸胃不適……
不過把雜果賓治撇開後,能夠在同一個地方幾天內吃到各國的美食也算是不錯……(你說甚麼?香港也能吃到上述各國地道食物?這點我也不得不承認,但我在這裡是衷心覺得他們安排不錯的呢。)

這晚沒有英超,我們卻再次聚到同一開房裡開通宵,將那萬惡的簡報預備好,然後爭取為數不多的寶貴休息時間。
……

Saturday, 27 November 2010

偽.SIMC回憶錄3.1

6:45AM, NUS High School
我們一如既往來到這裡吃早餐,今日的早餐是泰式年糕還有三絲炒米(當然還是香港的好吃XD),可是我們明顯沒有心情去大快朵頤,因為很快題目就要正式公佈了。
公佈題目的儀式來得簡單而隆重,只是上台領取一份題目和一隻光碟(Softcopy),確認題目就可以開始了。每間學校會分到一間班房,我們學校有兩隊,兩隊在同一個班房。
題目大致上是這樣的:
香城(?)有2間咖啡連鎖店,分別是Starbucks和Coffee Bean(為甚麼不是Pacific coffee? XD)
前提:人口平均分佈;等比於街道長度,另外一條街的market share會由最近這街的咖啡店拿到,如果一樣近則平分market share.
Q1 找出最好的開店策略,並在給定的地圖中找出用最少的店做出最高market share
Q2 給定一個資金,假沒開店費正比於街口的路口數目,用給定的資金做出最高market share
Q3 建立一些更真實的模型

解題目的思路等的的或許留到後面再說,以下是解題以外的花絮:
-要佔領全市的market share?很簡單,每個街口開一間Coffee Beans!
-我們計了整個上午後,玩了一會Osu!作為休息,以為可能會導致題目做不完,結果我們經過其他課室時發現,我們未必是最專心的一隊,但一定不是最hea的一隊。一牆之隔的中國隊已經在睡覺聽歌;某外國人更在下面花園水池站在池中央的石頭上拿著問題擺甫士沈思!
-眾所周知,那邊的天氣頗熱,於是電腦不停出現過熱的情況,我們試過將支裝水放在電散熱位附近,過了一小時,瓶蓋因為裡面有水蒸發,壓力太大而噴了出來!(此為危險動作,請勿在家中模仿)
-跟我們一起的那幾個星洲學生因為太無聊於是自成一角打起機來。他們在打GBA。嗯,是GBA無誤。而且是在玩遊戲王 XD
-入夜後大夥已經擬定好通宵的工作,夜上就變成休息時間了。大家主要都在打撲克,從鋤Dee到橋牌都有……

10:00PM, Day 2, 宿舍
晚上的工作似乎就是大家將早上分開做的題目合起來,做個結論,打一下報告文本而已。大家之前熬夜經驗不多,我們手上只有2部電腦,第三個人通常就在睡了。還好,之前買的咖啡和荼成為了各人的提神劑,讓大家也能勉強做下去。最神奇是……我們是被球賽叫醒的。宿舍的網絡很快,我們就一邊接上了英超直播一邊做,結果球賽開始時大家居然不約而同起醒來了。
然而,當晚要做的事總算做完了。

Thursday, 28 October 2010

偽.SIMC回憶錄2.3

6PM, Day 1 @ Singapore
晚上的行程其實也不多,也就是吃個晚餐多一點而已。沒有豪華的餐館,我們到了一間普通的Food Court,吃的更能夠反映出地道的食物特色。我點了個海南雞飯,質素應該比香港不少茶餐廳好。
再搭過一個地鐵站(順帶一提,那邊的地鐵也叫MTR,裡面的結構跟東鐵線沿線各站的建築十分相似)來到那邊的大會堂(真實名字我已經忘掉了,但其作用就是大會堂,上演演奏、歌劇這樣)。我們去到的時候一隊rock band正在演奏著,考慮到當時是週一晚上,觀眾不算多,但每個都是帶著滿腔熱血的粉絲,這跟香港形成了鮮明的對比。為甚麼香港非主流的音樂被邊緣化的那麼厲害?想聽的人不是沒有,但是怎麼他們難以出頭?
……
在大會堂那邊等了整個小時旅遊巴,我們總算回到宿舍,領了些免費物資如水後我們各自回到房裡去。在這裡支裝水上面貼著的標籤值得分享一下:除了NUS High School以外還有Experiment, Explore; Excel這三個詞語,正是Experior, Expono, Excedo這個口號的英文翻譯。數學不單是學習理論和應用,更重要的是推廣和一般化來解決更深層的理論,而實驗(失敗與嘗試)、發現和推廣正是研究的不二法門。但香港的教育制度下又有多少人是這樣看待數學的?……
此時接上網的電腦MSN開始被轟炸,我也沒能力逐一回覆,草草看完新聞就睡覺去。
……
新加坡有沒有葡萄適?
……zzz
12AM, Day 2 =END DAY=

Tuesday, 26 October 2010

偽.SIMC回憶錄2.2

(上次那篇是本blog的第500篇文 0w0)

海報展示就是每隊將自己預先製好的海報展示出來,一張介紹自己的學校(的數學活動),一張介紹一位本地數學家。這時將會有不同老師和校外來看不同學校的海報,各隊成員也不例外。
要在這裡逐一介紹不同的學校海報實在不易,但有幾張對我來說比較深刻:中國某隊將CMO列為學校數學活動,更令人驚嘆的是該圖片考生拿著的考卷似乎是高考卷!另外一隊中國隊用我們在地鐵站外經常看到"$30/1200分鐘"這樣的直立海報而不是標準的A2大小……果然是面子工程?另外某日本隊似乎懶得寫活動上去,只是出了道數學題,卻也吸引了不少人。
我們的海報沒啥特別的地方,活動是看似運動多於計算的Treasure Hunt。這是最重要的人物出現了--比賽的主裁,也是香港人的Raymond Chan。身為香港人他也作出了最低限度的聊天(避嫌?),同時也給予了我們不少鼓勵。
……
正式的比賽始於第二天,第一天當然免不了有破冰活動,原來世界各地的破冰活動基本上都一樣 XDDD
接下來的半天是促進旅客消費(?)的活動,我們分成不同的組別(每一隊都被打散開去,一組十人都來自不同國家)去遊覽星洲幾個景點(其實也不算"景點" orz)
第一站是China Town,故名思義裡面的人都會說中文(?)、賣國貨(?!?)還有……到處都是繁體中文!(感動)我在那邊買了把扇(可惜已經不見了),上面寫著橋夜泊,但並不是以七言的格式寫成,在此分享一下:
月落烏蹄 霜
滿天江楓漁 火
對愁眠姑蘇.城
外寒山寺夜 半
鐘聲到客 船
這樣讀法似有言外之意,但又不太通,到現在還沒解的懂……
然後我們到了一個類似香港熟食市場的地方吃紅豆冰,膠面的桌椅、灰色的牆漆、濕漉漉的地板,這一剎那我彷彿回到了香港,可惜我面前的是紅豆冰而不是腸粉。
我們沒有乘搭旅遊巴,而是選擇了當地的巴士。當地的巴士是以區域收費,而不是香港那種從總站算起的分段收費。同時他們也在用一種跟八達通差不多的卡,現時只能在交通工具使用,但是據他們說也在普及化了。
我們接著來到一個跳蚤市場(要跟香港比的話就是聯合廣場,跟信和比不上?),裡面有著新加坡各種潮物。可是我們也沒啥興趣買那些東西,只好跟那邊的學生閑逛和閒聊一下。很快他們帶我來到一間書店面前。以往的經驗告訴我們那邊的書店以英文為主,當然也不合我口味。但是當他們把我推進去的時候,我赫現發現裡面居然放滿了(繁中)日本輕小說!這也太變態了吧XD

Tuesday, 12 October 2010

偽.SIMC回憶錄2.1

Day 1, 5:45 am @ Singapore
沒想過會這麼早被敲門吵醒。
敲門的是鄰房的同學(似乎房號是隨機抽的,但是我們很幸運地中了鄰房XD),我矇著眼踱過去開門,門外的同學當然也沒睡夠(當然了……睡眠時間才3小時 orz),可以一言驚醒"夢中人":「今天開往主辦場地的巴士6:30am開,逾時不候喔。」睡意在當下就被消去得一乾二淨。經過一輪執拾後我們總算左六時半前出發,隊友當然也免不了掛著黑眼圈。
這是我們第一次前往這間學校,一如之前的的想法,我並不認為這所學校有多大,所謂的On-site challenge也大概是IMO那種吧。結果居然出乎我的想象。

我們的早餐當然在飯堂中完成,似乎是出自主辦單位的意思,我們接下來每餐的食物都有著不同地方的特色。最令人意外的是這天竟然是「港式」早餐!好吧我承認我誇張了。這餐基本是西多士、火腿煎蛋通粉之類的食物。(外國人對火腿配煎蛋感到新奇?)最特別的要數就是飲料了。基本上每餐我們都有三種飲料選擇:雜果賓治、(紅)荼和咖啡。
很不幸地,那該死的賓治是藥水般的桃紅色,理所當然地也散發出櫻桃的香味。第一次在這裡的餐的我們當然不會知道,可是這玩意我們飲了一次後卻也沒有抗拒,每餐總來一杯。有趣的是,到了後面幾天,這裡提供的雜果賓治變得越來越淡 Orz 此外,這杯雜果賓治被懷疑帶有瀉藥作用,這些容後再談。

六時四十五分抵達學校,我們足足有75分鐘用餐(當然我寧願將其中60分鐘留在宿舍睡)。平時習慣秒殺早餐的我們當然有不少時間剩。在這空餘時間裡面我們拿出了手提電腦玩。重點是,我們看到類桌的中國某隊跟著我們拿出手電玩,玩的遊戲居然是LF2 (Little Fighter2)!這也算是HK青少年的集體回憶吧~還好他沒有玩GBA模擬器,要不然我大概就笑出來了 XD

我們就一直玩到八時,聽了一個簡介,接下來就是展示學校海報的時間。
……

Saturday, 25 September 2010

偽.SIMC回憶錄1.3

手拿著從同學借來的手電,到宿舍後的第一件事當然就是把電腦"settle down"--比如檢查郵件之類的。這時我注意到桌面上放了一個黑色背包,背面寫著SIMC的字樣。好奇心驅使我打開了那個背包(用另一個說法:那個房間"是我"的話,那個背包不屬於房間的"應有設備",就一定"是我"的了。)
裡面裝的物品也不算出乎意料之外:一個萬用插頭(當然寫著SIMC, 從插口和"腳"基本上萬能,看來不只在新加坡才能用);一本SIMC簡介(只能用精美來形容,連各隊的學校簡介都有);還有一個放著NUS(新加坡國立大學)的資料,最神奇的是裡面只有應用科學那邊的資科(純數啊純理化那樣都沒有);一個印了logo的水壺,800mL,相當實用;當然還有之前訂了的SIMC T-shirt (前後都印著logo、用運動用布料做的T-shirt)。說到這裡不能不讚一下新加坡那邊的預備相當充足(這也是無可厚非的,畢竟預備了一整年),幾乎所有要用的物品都齊全了(就差一本簿和筆? XD)
正式活動是在Day 1開始,當中不少隊伍的航班都比較晚,於是這天晚上成為了(唯二的)自由時間。
經過一輪商議,我們最終決定了由幾位同學"帶隊"去四處遊覽一下。
遊覽的地方我就就不多說了,主要是聖淘沙(看Songs of the sea, 水柱只能令人想起Waves),值得一提的是看來當地小吃比HK便宜:以1兌5來算的話,$1.4有杯紅豆冰或者一個格仔餅的小吃,7元在HK只夠買一串串燒啊~)
當地的紅豆冰概念跟香港的牛雜都差不多(大誤),就是將每種配料都加點進去(當然紅豆冰會以紅豆和冰為主),就我所看見店主先後放了紅豆、椰果、冰、菠蘿,再淋上N種醬汁(我能分辨出的就只有花奶、煉奶和可樂(對,是可樂啊啊啊)還有紅黃綠三色糖水。這樣的食物在香港我大概絕不會吃,但是在那邊我意然沒經大腦思考就整個吃掉了 OTL
另一件事是當地都有類似日O城十元店的東西。十元兌過去就是每件貨品2元,於是我就捧了二支1.5L的茶回去了。

宿舍還是老樣子,只不過不少其他隊伍也陸續到達了。我們各自離去回我們自己的房。這時最離奇的事發生了:我和幾個同學的房間都在同一棟,但去不同樓層都需要自己的卡匙,例如五樓的卡匙就不能乘電梯到六樓。其保安之嚴密,在每個樓層走廊的門口都另有一個要卡匙進入的大閘。這次有個同學按了八樓,住六樓的我卻來不及按六字,電梯自然直上八樓。我心想:八樓的話用走火通道下去就沒問題了吧?我打開防火門,卻沒有發現這是向外反鎖的門,到我跑到六樓氣喘吁吁地求救時已經太遲了。我只好跑到地下的防火門,連地下的防火門都鎖著!(說實話地下不鎖 其他樓層鎖就可以了……)還好大力拍門的我最後還是獲救。

回到宿舍要睡時已經是凌晨兩點了。我實在等不及明天的挑戰……

Friday, 10 September 2010

偽.SIMC回憶錄1.2

Day 0 12:00, SG
機上的四小時完全是在紙堆中渡過。我不停翻讀著那份要命的筆記……還有那一本忍不了手買下的的經濟學概論。所幸的是機上有午餐時間,我也不至於專住了好幾個小時。
降落新加坡後,第一件事……居然是去廁所(XD)~在此之前我一直認為新加坡在那種管治下不論是文化或是教養方面都有一定質素,可是這一下正好推翻了我的想法。
明明是一個機場的廁所,裡面跟香港的公廁沒兩樣,實在慘不忍睹。
至底這是個區有金玉其外而敗絮其中,還是表裡如一的城市?
這個問題成為了我此行的另一個目的。
辦完手續後當然就是正式進入新加坡了……沒想到另一個驚喜正在等候我們。
*_*
一個月前當我們接到比賽的簡介時,單張上寫著的是由新加坡國立大學附中所舉辦的比賽,當時我就想:「有沒有搞錯啊~一間學校辦的比賽認授性本來就不足,何來『國際』?」於是當時就抱著玩玩去的心態報名去了。
甫踏進接機大堂,陣陣鎂光燈向我們閃爍著,我第一個反應便是「咦,狗仔隊拍錯目標了?我又長得不帥~」於是抬頭一看,赫然發現幾個當地的同學拿著「xxx school welcome to Singapore!~」的橫幅!
嘖,錢不少呢。
相機閃夠了後我們自然走了出來,三位年紀跟我們相若前來自我介紹,一共是兩男一女--為了不公開名字,我們即管把男同學稱為A,B,女同學稱為C吧。在一番談話中我們得知他們會負責領隊的角色,在這幾天跟著我們。
在談話中實在不難理解他們被挑選成為student leader的原因:在旅遊巴裡面沒有冷場,除了是因為有英文和普通話這兩種共通語言外,還因為那三個同學的熱情和健談。旅遊巴經過所見到的景點,從雙子塔到法院大樓他們都沒有放過。有了建築特色這種兩地的共通點,各人很快就能融入大家的對話當中。
新加坡的面積比香港小,車程不到半小時,我們便到達了以後五天的住宿地方。當旅遊巴在宿舍外圍繞圈時,我還以為是豪宅之類的建築,但竟然只是大學宿舍--那是山上劃出來的一塊不少的區域,上面矗立了十餘座約十樓高的建築物,外面看起來由落地玻璃砌出來大樓,實在很難想像成永遠和鬼故的大學宿舍扯在一起。
宿舍房間的感覺截然不同,說白了就是一種實而不華的設計:一張床、書桌、景觀還不錯的窗口和浴室;整個房間漆上柔和的淡黃色--宿舍,本該如此。

(待續)

*嘛, 我懶了...很多東西都沒寫出來呢

Thursday, 12 August 2010

偽.SIMC回憶錄1.1

之前那篇英文的大致上都是有關題目剖析的內容,至於這篇,當然希望是一篇堅的回憶錄了。它之所以是偽回憶錄,是因為文中的人不一定是"我"、卻一定是屬於過去、我的回憶所見所聞。
"Those were the days my friends..."
時光回到四年前.英國,倫敦。
「Nice to meet all of you and the bygone time will be our precious memory forever. Now, shall we take a photo?」一句話,一張相片,將一瞬間封印成永恆。不過,一與零的組合真的可以成為永遠的回憶嗎?
數不清的教訓告訴我們,世上哪有這麼便宜的事,一張照片就能將過去繪聲繪影地呈現在記憶之海中?
一個個零碎的記憶、恍如羽毛般飄走的追憶……那被封印的完整瞬間,卻成了一絲遺憾。
四年後,香港。
這是一個沒有假的暑假。書展、動漫節、然後是數不清的補課。正當我們放眼將來時,另一些事物卻悄悄飛逝,成為茫茫歷史中的一份子:會考放榜標誌著英式3223(啊,不是足球~)學制的終結……
始終我是334的人,也沒有玩過會考,當然對會考也沒有這樣的情懷,可是……
將各種碎片拼合起來,構成的便是人生。大塊的碎片較顯眼,沒有這麼輕易被遺忘;小塊的是日常的細節,卻占了拼圖的一大部分。
一張照片可以封印一個瞬間,一段文字卻可以封印一段時空,將彼岸的記憶呈現過來,回味那珍貴的回憶。
Day 0, HKIA
幾個少年從Axx號巴士徐徐走出來。在他們身後的是一個個大得嚇人的行李匣。現在明明是五月中,考試也即將來臨,根本不可能在這種時候外出旅遊,更別說拖著這樣一個大行李箱了。
他們隨著人流走進離境大樓,迎接他們的是另一群包括了我少年。
「真準時呢。」我對他們打招呼說道。
「這種『有今生、冇來世』的盛事當然要準時嘍。再說,昨晚也不可能睡得著吧?」
「哈,昨晚有足總盃決賽,我看完就動身過來了。頂多就在機上睡一睡吧!」
「機上睡……可是我還沒有看過那一份資料呢?」
Network Optimization那份?那已經不像是中學能看懂的東西了吧?一堆矩陣和複雜的計算,我都不想看了耶。」
「可是去年的題目也是這樣,出了一堆複雜到爆的計算,不用矩陣可能計到虛脫吧?」
「這個不用擔心啦……今年賽制變成36小時生死戰,數據一定比去年簡單呢。」
「唔,說得也是。」
「既然這樣,大家就先去吃個早餐好了。一會在這裡見囉!」
吃過早餐的我當然不會選擇呆等,而是回歸我本身的嗜好~看書。碰巧我附近就有一間Book One,我毫不猶疑地走進了書店。
隨手翻起書架上的一本書,正是一本經濟概論:
「在一個給定的流程表(flow diagram)下,其對應矩陣A在滿足AX為一極端值時,其運行效率亦會提到最高。」
……

Wednesday, 2 June 2010

SIMC report (5) Challenge 3iii

7)Long term consideration
In model 1 (Challenge 1), our locating strategy is about to miximize the market share with the least number of new shops opened. That is, only consider the marginal influence of the shops.
In model 2 (Challenge 2), our locating strategy is about to maximize the market share within the given amount of budget, proportional to the k-junction. That is, the cost of opening cost is the only new factor.
Now let's do some math. Assume the cost for the n-th month is C(k). Generally C(k) is increasing due to inflation. Extend C(k) so that the domain the real number. (negative k implies that cost projected in the past), also C(k)≦C([k])≦C(k+1)  for any integer k. Therefore it is monotonically increasing. Integrate once on C(k) from zero to infinity, which is diverging. Therefore the cost of running a shop under unlimited time requires infinity cost. Comparing with the cost of opening a new shop, which is a finite cost, the ratio of run cost is far (infinity-ly) more important than the set-up cost. So in case we should take the maximization of profit as the second aim after the maximization of market share. Indeed, what we have to do is build up a model to consider the lattice that supplies. To extend this model, the vector field of supplier and consumer can be built, which will be the ultimate model of this problem.
"We can't open the shop under an unlimited time. Why we have to consider this case?"
This assumption helps us to realize that the running cost is more important than the set-up cost in running a shop, which will be more important than the locating strategy to win the competition.
"If the maximization of profit is concerned, using multiple shops to block the Starbucks will be useless. Then all previous models will be demolished."
Yes this is the main problem of the simplified models.
8)Debug
So far we have finished the analysis of this problem. You may ask that why we finish this problem with just some simple maths, but that is the truth. In fact, our proof about the locating strategy is incomplete, but in the sense of considering every simple case, it works. Combining together, it will works too.
Some advanced tools like matrix and programmiung is  highly appreciated in the competition but we tried to finish the problem by simple mathematics, we will be proud of this.
The full solution from different team will be released in 1 month and I'm looking for this and a further analysis.

End of SIMC report: The Challenge

Tuesday, 1 June 2010

SIMC Report (4) Challenge 3ii

ii)A complicated model
When we built up the previous model, we assume that the population only flow within the two closest shop. In fact the population flows everywhere. That how can we express the average population density as a average of a whole day? Putting them in equal proportion is obviously not fair. Our solution will be:
Assume that for a particular street, the closeset distance from the street to the Starbucks shops willbe a1,a2,..., am units while the distance to the Coffee Bean will be b1,b2,...,bn units. Then our market share will be: [Σ(1/bi)/Σ(1/bi)+Σ(1/ai)]. The idea is that we still have to show that the closer distance will have a higher chance of providing service to a (averagely) closer customers.  It can also shows that a very close shop does have a very strong influence on that street (which is the real case).
"Distance tends to zero will results in an infinity-ly large influence to the streer?"
This problem makes no sense. The distance is assumes to be an integer, and the indfluence is measured in terms of the whole street. Even though the man living besides the shop goes there everyday, the shop can't influence the whole street at all.
Now note that consider the model when there's only 1 Starbucks and 1 Coffee Bean shop. The market share will be (1/b)/(1/a+1/b)=a/(a+b), which is just same ax the model of (i)! Then we are success in considering the extended model.
in this model, the locating strategy is nearly the same as (i) so we need not to discuss again. Just a point to note that the locating of n-junction will be much more important.
One important function of this model is that the share of a shop to a particular segment of a street is not as exact as 1 or 0 or 0.5, but they're equalizing to half-half. They are competing. That's why we need some strategy to win the competition.
(Note that the share originally is 1, falled to less than one, higher than 0.5, if the share is half-half it's still the same, and if it is zero it will raise a bit.)
6)Population denstiy
In the above model we are considering the flow of population. However one of the assumption still exist -- the population density is even. If it is not even, the proportion will be different.
In the real society, Coffee Shop starts their business in the business zone where people used to negotiate there. A rural area will have a higher potential population density than a urban area. It will be very complicated for us to compute those values. In order to show that the model is efficient to compute the real model.
consider the above map (2*4) again. Now assume that the segment (1,2) to (3,2) is a trade zone where people likes coffee so much there (they don't put a coffee machine at office right?), hence a double share to the coffee shop. The new share will be 5.5/16, where the rate of increasing will be 14% which compansated the deficit of locating in a bad location which fewer customers.
Combining thw two models, the new share will be about 7.2/16=0.45 which is nearly half! We see that the effect of trade zone is even larger than the density.

we will take care of the long-run considerationg in the next discussion.

Monday, 31 May 2010

SIMC Report (3) Challenge 3i

In challenge 3, we are required to create a few models considering real factors.
5)Potential Population Density
Recall the assumption that have been made in Challenge 1,2: They go to the closest shop only; continous and even distribution of potential population and a given map of street.
What can you see that assumption different from the reality?
Going to the closest shop implies that the potential customers are in stationery. In the real society, the population forms a vector field (on a 2D/3D coordinate, whatever), where the peference of cumtomers changes from time to time. For example, we won't go to starbucks in the early morning. Therefore a multiplyer t(x|24) is set to be a cycle of population density in different time.
Consider the given potential population density at a given instant. The vector field of the customers can be ignored since the vector field indicates the rate of change in density and indicates the density in the next moment only. At a given moment there're no population flow. Now the question is: it is impossible for us to create a model and a vector field to compare the marginal influence at every instant. How can we create a easy and reasonable model?
Our solutions will be:
i)A simple model: Consider a street, apart from the closest Starbucks and Coffee Bean from a and b units, then the share of Coffee Bean will be a/(a+b) units. The idea comes from that the closer shops will have a higher probability of selling services to the closer customers. The assumption is that the population is flowing, but there will be a higher chance to walk along the closer shops.
Consider the model on a 2D- plane: (0,0)to (4,0), (0,2) to (4,2), (0,0) to (0,2), (2,0) to (2,2), (4,0) to (4,2), total 14 untis. Assume that the Starbucks located at (2,2) while the Coffee Bean is located at (0,0).
It seems that the location of Starbucks will be much better than the Coffee Bean. Actually in the original model we will get the share as 5/15 units at the proportion of 5:9. But it  is quite weird as they are close to each other. Then how can we improve the result?
In this model, the new share of Coffee Bean will be 6/14 (6:8) which is more reasonable.
We can see that in this model the strategy of "blocking others" is not effective anymore since two close shop will have similar shraes. The new locating strategy is main locating them in the junction point and near to the centre of map which is more close to reality.

Unfinished and we'll discuss a complicated model considering the population flow tomorrow.

Sunday, 30 May 2010

SIMC 2010 Report (2): Challenge 2

The aim of tis problem is to maximize the market share within a given budget relating to the position of the shops. Note that the cost is not related to the marginal influence, but the n-junction road,
3)Basic Strategy
Of course we should we most of the budget.
Proof: setting up a new shop shop increase market, so putting more is better.
Another important strategy is that we should bulid more shops in the 2-junction road.
Proof: it's the matter of blocking others and obtaining more "weak" controlling area.
Why obtaining more "weak" area is better? A basic frame is that the area of "strong controlling" is almost the same for every shops. It means that if you bulid a shop in their "strong controlling" area, you will get a little share. Then the "weak" area will be relatively bigger. Now when the number of shops is equal, the brand who controls more "weak" area wins. When three shops nearby forms a close triangle, the area within these three shops is somehow likely a "strong controlled" area so that a large area can be obtained.
comparing the 4-2, buliding two with 2 units of cost is obviously better than buliding one shop in the 4-junction road (even two besides the junction must be better). But how about the 3-2 comparason? It seems that the case of 3-2 is much more complicated since there're too many junction to be drawn and the stright road is undetermined. However we simplify the case for putting two at the end of the junction and the last one in the end of one of the junction. Then the share is also miximized.
Therefore our conclusion about the strategy will be maximized when we bulid as much at 2-junction as possible.
4)Application
Now we go the the core idea of the problem. How can we apply the strategy?
our solution is putting them in front of Starbucks to block them, and we get 27.5/47. however we can see that the optimal solution will be 28/47 where two blocks starbucks below and one above the 4-junction road.
Now we realize that the solution aimed to rule the whole street in the bottom. It seems that it's still the same target: to block Starbucks and acquire a large area. But what's the difference?
The difference is that we are trying to block Starbucks in their "strong" controlling area, and to obtain the whole street in the center. In their solution they tried to obtian the weak area.
Note that when we block starbucks, we can only block one of their way, but they still control the other ways. But acquiring the whole "weak" area is different from that. It will be impossible for Starbucks to go inside the "controlling triangle" such that we can surely win the competition.
Locating near the 40junction road aims at acquiring more market share from different road, then they will have the higher chance the control a "weaker" area. In our solution,
Recall the two strategy : near to the junction road and vlocking others, they have different advantage so how to balance them and maximize the share? That will be the key idea of our report.
Anyway I think this problem is just a further application of Q1, but not very hard if we have solved Q1 in a good way (or many infinite times of trial and of course error?)

SIMC report (1): challenge

For the quesiton please refer to this link.
http://www.highsch.nus.edu.sg/SIMC2010/Challenge%20Question%20pdf.htm
Firstly I would like to talk about this problem first since this will be the main propose for us to go to Singapore~
1)Basic analysis and definition
We now simplify the expressions: for X----O with length L, the market share of Coffee Bean (O) will be L/2. Sum of every line segment between two closest Starbucks shop/Cofee Bean shop will be the total share of the whole map.
2)strategy and algorithm
What is the difference of strategy and algorithm? In this problem it's just the same, but dirrerent in score. We have outlined a several strategy (not absolute rule), but we can't obtain enough score here. The main reason will be the difference in complexity and the ensurance of optimal solution.
i)Disperse without overlapping
Proof: Consider X--O1-----O2, XO1= k units, XO2=L units, then market share will be L-k/2 units.
Therefore as O1O2 disperse (L-k increases), our market share increases.
ii)Near to junction road
Proof by a 4-junction road. Consider a CB shop near to a cross junction while it market share to the three road will be a,b,c units respectively. If CB shop is 1 units near to the 4 junction road, the market to each road increases 0.5 each which is benificial to CB shop.
iii)Blocking Starbucks shop
Proof: consider the model in (i) again: minimize k to 1 (k≠0, k is non-negative integer), we maximize l(x)=L-k/2.
That's our work to find out the optimal solution.
Greedy algorithm
Marginal influence of a lattice is defined as the increase in market share when a new shop is added there.
Method I: Put the shops in the highest marginal influence.
Method II: Put one shop in the optimal place, re-calculate the marginal influence and to place again.
Obviously Method II is better since optimal three ≠ best combination.
However, we can't show that method II is optimal since it still just a combination of optimal shops, but not an optimal combination.
"I know your method helps to find a better solution. But how can you find such a solution. I mean, is it a modified trial and error?"
Modified trial and error is not avoidable at all. Even a full trial and error will be acceptable. We will that according to our strategy, a few choices will be left as so that we can try all of them.
"How can you determine the optimal solution among them?"
this one is stupid, just by trial and error and comparason.
"Blocking Starbucks is unfeasible. They still have another way to control their markey share. What's the main effect of this strategy?"
First we define "weak" and "strong" of a control of a shop on a particular street. It will be relative concept that is proportional to the distance to the closest shop. Placing a new shop in a "weaker" segments will have a higher market influence.
If we block them is the same direction, the opposite-blocked direction will be "far" from Starbucks so that Even the CB shop is controlling some area somehow "weak",  Then a large piece of area will be in the contrl of CB.
We don't have much to discuss with the solution of (1b), it's just a direct application of our strategy.
4 junction road + blocking the SB on the LHS. The second one located just below the bottom-right SB.
Hmmm, about the City Hall map it's a true hard one. The map is complex and there're far too much combination for us to call it "we 'have fully tried all"....
Let's finish the report for Question 1 here.