二等獎

用塑膠軟管作為氫爆安全反應器，從封閉端點燃氫氧預混氣，管中火焰\r 傳遞快速，肉眼難以觀察，因此研製「氣爆顯影記錄器」來觀測，如下圖，\r 光纖導引偵測點的火光至暗箱，記錄於高速旋轉的軟片，從光電轉速計讀取\r 轉速，軟片沖洗後量測各顯影間距，可得火焰位置與時間的關係圖，量測時\r 間最小刻度可達2. 2 微秒。\r 解讀軟片的顯影，得知氫爆初期，火焰加速傳遞，大部分在25cm 位置\r 左右，火焰的速率出現急速飆升(爆燃轉變為爆震)，而且火光亮度也急速升\r 高，有時火焰速率出現飆升過高，再回降趨於穩定?的現象，在50 cm 位置\r 左右，火焰進入等速傳遞階段，此階段有穩定的火焰模型，火焰頂端的亮度\r 最高，往後亮度遞減。

在一個有障礙的平面上，給三個定點，我們探討連接此三點的最佳網絡。我們討論了諸如直線、射線、線段、圓、網格狀、三角形……等類的障礙，當網絡每穿越障礙一次，就必須付出代價，例如「拖延5 分鐘」。所以，設網絡穿越障礙的次數為y ，則網絡除了原本的總長度之外，還額外加入y 倍某固定數值的損耗。我們以費瑪點的各種性質及三角形不等式等方法為工具，就不同的穿越障礙次數綜合比較，而找出最佳網絡。在某些情況下，最佳網絡不是以費瑪點來連接三點，而是在障礙(如：直線)上找出符合某種與餘弦值相關特殊性質的點，以該點來連接三點，而此網絡可用GSP 軟體相當精確地作出。另外，我們也探討在考慮障礙造成損耗的情況下，兩點間的「實際距離」為何。 最後，我們考慮「混合障礙」問題。在此類問題中，除了前面所討論的障礙，還另加了如同「河流」的兩平行直線間區域之障礙，在這種障礙區域中，網絡的長度要乘以數倍來計算。我們發現，此類問題的最佳網絡也可用特定的正弦條件配合GSP 而相當精確地作出來。;Considering various kinds of obstacles in a plane, such as a line, a segment, a ray, a circle, a triangle or chessboard grids, which function like a red light, we research into the problem of finding the optimal network connecting three given points A, B, C in the plane amidst obstacles described above. Each time when the network crosses an obstacle, it will cause losses, such as five minute’s delay or a loss of one hundred dollars. Taking advantage of Fermat points, some basic inequalities concerning triangles and some special qualities about sine or cosine functions, we obtain the optimal networks in different situations. Besides, we consider what the “real distance” between two points is when there are obstacles in a plane. We also put another obstacle, including a line and a weighted region between two parallel lines, into consideration. In the region, like a river or a muddy ground in real life, the length of the network should be multiplied by a fixed time. Furthermore, we can use GSP to make the networks very accurately.

本研究以網路上Soft X-ray、UV195Å、K-Line、Ｈα波段太陽影像上黑子，光斑經度位置的逐時變量，試求太陽自光球層以上各層是否有越高層轉速越慢的現象？經試用數種方法求大量黑子、光斑位置，最後以省時，且誤差減小的＂由PhotoShop軟體讀出黑子、光斑在圖檔中位置，於EXCEL軟體中套用日面座標公式＂，以Peter Meadows的黑子定位程式抽樣對驗，來求其經緯度值。結果：在任何緯度，光球層以上各層轉速不同（徑向較差自轉現象），但並非越高層越慢，而是在大部分緯度處其各層速度變化呈 的快慢交替現象。

安非他命的濫用在台灣是非常嚴重的公眾健康及社會問題。安非他命會導致一連串的行為異常，包括在中腦紋狀體內釋放多巴胺及阻止多巴胺回收來增加使用者的活動力。由於安非他命會對腦細胞造成傷害，本研究的目的為探討低劑量、無立即毒性之安非他命(類似於人類使用習慣)長期施打下，是否會對C57BL6 小鼠大腦紋狀體內的蛋白質表現有影響。因此利用西方點墨法分析施打低劑量安非他命(2 到6 mg/kg) 約一星期之後，C57BL6 小鼠的大腦紋狀體中一些重要蛋白質(包括腺.酸受體A2A-R、第五亞型腺.酸環化.AC5、caspase-8 及PARP) 的表現是否有改變。實驗結果顯示，低劑量安非他命處理對這些蛋白質的表現並沒有明顯的差異。但利用二維電泳法可看到有少許蛋白質，在經過安非他命處理下有顯著的差別，如KIAA0193 homolog 、GOS-28、gammacrystallin A、malate dehydrogenase 和phosphoglycerate mutase isozyme B (PGAM-B)。這些蛋白質中，malate dehydrogenase 和PGAM-B 與代謝和產生ATP 有關，但前者是增加的，而後者減少，推測安非他命會影響神經細胞的能量代謝，因此長期施打安非他命對紋狀體造成的影響值得進一步探討。;The wide spreading use of amphetamine (AMPH) in Taiwan has become a serious public health and social problem. AMPH evokes a series of behavior abnormality including enhanced locomotor behavior by releasing dopamine and inhibiting dopamine-uptake in the striatum. Since AMPH is known to cause brain damage, the purpose of this study is to investigate the expression of several important proteins in the striatum of C57BL6 mice after chronic treatment with low and non-toxic dosages of AMPH (mimicking the common usage pattern of AMPH addict). C57BL6 mice were daily IP-injected with various dosages of AMPH (0 to 6 mg/kg) for one week. Expression levels of A2A adenosine receptor (A2A-R), adenylyl cyclase type V (AC5), caspase-8 and PARP in the striatum were analyzed by Western blotting analysis. Most proteins examined were not affected by this 1-week AMPH treatment. By the aid of two-dimensional gel electrophoresis, expressions of a few striatal proteins (such as KIAA0193 homolog, GOS-28, gammacrystallin A, malate dehydrogenase and phosphoglycerate mutase isozyme B (PGAM-B) in AMPH-treated mice were altered. Note that malate dehydrogenase and PGAM-B are two enzymes involved in energy metabolism and ATP generation. Interestingly, the former was increased and while the latter was decreased in AMPH-treated mice. Collectively AMPH may affect the energy metabolism in neuronal cells. These results suggest that the injury induced by long-term AMPH exposure warrants our further concerns and investigation.

本研究將探討所合成的金屬硫化物製備於量子點敏化太陽能電池之光電極，並應用於水裂解反應，使之產生氫氣作為新型替代能源。以硝酸鋅、硝酸鎘、氧化硒及硫化鈉作為起始物，利用化學浸泡法(Chemical Bath Deposition, CBD) 於氧化錫參雜氟之導電玻璃(fluorine-doped tin oxide, FTO)表面上，製備不同層數量子點硫化物之光電極。當CBD循環層數為1、5及7時，運用紫外光-可見光吸收光譜儀與掃描式電子顯微鏡分別可觀察到光譜變化，與球狀、團簇狀，以及緞帶狀奈米材料之結構。另外以線性掃瞄伏安法測量光電流，在一個太陽光強度(100 mW cm-2)照射下，當電壓為1.3 V (vs. Hg/Hg2Cl2) 時， TiO2/(CdZnS)4/(CdZnSe)2/ZnS光電極之光電流可達 79.24 mA，可證明經CBD循環製成之量子點電極具有高電催化活性。此外將沉積不同量子點的光電極與硫化鈷反電極結合後製成不同的太陽能電池，在一個太陽光強度照射下測量其光電轉化效率，當使用的光電極為 TiO2/(CdZnS)4/(CdZnSe)2/ZnS 時，可得最大轉化效率η為 2.40 %。更進一步，以此光電極在定電壓下進行水裂解反應，由自製的氫氣收集器收集氫氣，可得其與白金對電極產氫率為 0.98 mmol/hr，以此可確定最佳化的光電極。

在利用電腦螢幕來瀏覽圖形式資訊的時候，常常受限於螢幕的空間，沒有辦法在顯示\r 資訊整體結構的同時顯現細節部分的資料，目前的使用者介面所採用的方法有放大（zoom\r in）、捲動（scrolling）、開啟多個視窗（multiple view）等方法，這些方法雖然可以呈現出資\r 料的細節部分，但是仍有其個別的缺點存在，放大的方式會有遮蔽的情形；捲動的方式無\r 法同時地呈現整體結構；開啟多個視窗的方法使得使用者的眼睛必須在這些視窗間來回的\r 移動，造成麻煩。\r 魚眼鏡頭是一種短焦距、大視角的相機鏡頭，鏡頭成像的時候，越接近鏡頭中心的物\r 體會越放大而越遠的部分會越縮小，藉著發掘魚眼鏡頭的成像函數，我們發展出了一種新\r 的使用者介面，在瀏覽圖形式資訊的時候，能夠顯示整體的結構，並隨著滑鼠游標的移動，\r 以不開啟新視窗及無遮蔽的方式，即時地將想要觀察的部分局部放大以展現細部的資料，\r 這種使用者介面將具備現有方法的優點而無其缺點。\r Browsing the global structure of a large graph in limited screen space has the drawback that details\r are often too small to be seen. The most common solution provides a scrollable view. This shows full\r details at the region currently visible through the view, but hides the rest of the global structure.\r Alternatively, zooming into a part of the graph does show local details but misses the overall structure of\r the graph. The multiple views approach, one view of the entire graph and the other of a zoomed portion,\r has the advantage of seeing both local details and overall structure, but has the drawback that parts of the\r graph adjacent to the enlarged area are not visible at all in the enlarged views.\r A fisheye camera lens is a very wide angle lens that magnifies nearby objects while shrinking distant\r objects. It seems to be a tool for seeing local detail and global structure simultaneously. By means of\r exploring fisheye camera lens, we develop a user interface for browsing graphs using program analog of\r fisheye lens. Thus, our method seems to have all the advantages of the other approaches without suffering\r from any of the drawbacks.\r \r

「風笛」是台灣原住民鄒族的信號用具及祈雨法器，由一條繩子綁一支竹片構成。轉動風笛時，竹片會繞繩子自轉並拍打空氣而發出聲音，並有上下飛舞的現象。風笛產生聲音的原因，為竹片拍打空氣而造成的渦流共振現象；又由於繩子扭力大小及方向改變，使風笛的音調忽高忽低、響度忽大忽小、且竹片會在兩個平面上公轉，而有週期性變化。施力使風笛公轉轉速加快時，竹片自轉速率也變快，使其音調愈高、響度愈大；而繩愈短、愈粗時，竹片的公轉週期將愈短。The wind whistler was once used by Tsou aborigines as a tool for message transfer. It is composed of a string and a bamboo flapper. When swung around, the flapper spins, beats the air, and makes sounds. Moreover, the flapper flies up and down during the revolution. The spinning flapper beats the air, causes the vortex resonance phenomenon, and thus produces sound. As the twist torque and direction change, there is periodical variation in the sound volume, sound pitch, and the movement of the flapper, which orbits up and down at two planes. If given force to speed up its revolution, the flapper，s spinning frequency also increases, which makes the sound pitch higher and the sound volume greater. Besides, when the string is shorter or thicker, the flapper，s revolution period will be shorter.

Early physicists such as Newton thought that all objects have definite positions. For example, they thought that an apple is either inside a fruit bowl, or outside of it. The advent of quantum physics in the early 20th century proved this viewpoint wrong. There is an uncertainty in the position of any object; we can find a set of possible locations where the object might be. This concept was termed superposition. Quantum tic-tac-toe (QT3) elegantly extends the popular game of tic-tac-toe by adding this quantum physics concept of superposition. Each turn, 1 piece is simultaneously played into 2 distinct squares of a 3-by-3 grid. Eventually, however, every piece will occupy exactly one square, like in tic-tac-toe. Yet, despite this intriguing addition, not much research has been done on the game. Hence in this paper we explore the game in terms of extension, analysis and solution. Firstly, we note that the quantum extension proposed by Alan Goff in QT3 is incomplete. In reality, there can be more than 2 possible locations for any object. Unfortunately, the QT3 game rules do not allow for this extension. Thus we non-trivially generalize the game (GQT3) by proposing a new set of rules. We show that the original QT3 is a subset of GQT3 and prove that our generalized game can always be successfully played from start to finish in a finite number of moves. Then, we begin our analysis of GQT3. Firstly, we investigate the game tree complexity, state space complexity and computational complexity of the game; indicators of how complicated the game is. Notably, we find here that QT3 has a total of about 18 trillion possible games, which is substantially higher than tic-tac-toe’s 400 thousand. Then we examine the Nash Equilibrium of the game; the result if two ‘Gods’ play the game against each other. We find that in this scenario, the first player will win by 0.5 points. To make the game fairer, we suggest minor variations on the scoring, which make the Nash Equilibrium a draw. Note that standard methods to analyze all of these would take at least a year, but we bring down the time to about a minute using symmetry considerations and other optimizations. Finally, we extend our programs into an artificial intelligence that is a perfect solution to the game. We then supplement this with a utility function to make the run-time performance pragmatic for more time-consuming versions of GQT3. Ultimately, GQT3 is a challenging and unique game with myriads of exploration possibilities; we have only scratched the surface here.

The Boolean Satisfiability problem, called SAT for short, is the problem of determining if a set of constraints involving Boolean (True/False) variables can be simultaneously satisfied. SAT solvers have become an integral part in many computations that involve making choices subject to constraints, such as scheduling software, chip design, decision making for robots (and even Sudoku!). Given their practical applications, one question is when SAT problems become hard to solve. The problem difficulty depends on the constrainedness of the SAT instance, which is defined as the ratio of the number of constraints to the number of variables. Research in the early 90’s showed that SAT problems are easy to solve both when the constrainedness is low and when it is high, abruptly transitioning (“boiling over” ) from easy to hard in a very narrow region in the middle. My project is aimed at verifying this surprising finding. I wrote a basic SAT solver in Python and used it to solve a large number of randomly generated 3SAT problems with given level of constrainedness. My experimental results showed that the percentage of problems with satisfying assignment transitions sharply from 100% to 0% as constrainedness varies between 4 and 5. Right at this point, the time taken to solve the problems peaks sharply. Similar behavior also holds for 2SAT and 4SAT. Thus, SAT problems do seem to exhibit phase transition behavior; my experimental data supported my hypothesis.