| Abstract: | 區域規劃(Region Planning)長久以來都是生活上的重要課題，早在沒有電腦的時代這些課題還是需要以人工作業解決。時至今日，藉電腦的輔助區域規劃可以應用在學區劃分、便利商店展店分析、警車巡邏範圍、農地灌溉面積、都市計畫等領域。過去一些簡便的方法像是徐昇氏多邊形(Voronoi)等可提供不錯的規劃輔助，但是這些方法對於區域劃分都著眼於區域中心點位(Point)資訊，並沒有再加入其它空間條件設限，於是在某些領域的應用上便受侷限。舉例而言，某住家面臨百米路寬，對面就是一間小學A，但緊鄰隔壁五百公尺有另一間小學B，此區域到底該劃分為學區A還是學區B？以往常常都是先自動劃分為學區A，稍後因學童安全考量再人工微調至學區B。區域規劃對於一些大型屏障如鐵道、高速公路、山川大河等都該列入劃分時的考慮範圍，所以在此提出一個新的梯田演算法(Terrace Algorithm)可以自動考量服務中心與週遭環境，使挑選出來的答案能夠更切合生活課題。除此之外也提出一些評比原則來考量每個方法對於生活課題的適切性。就學區規劃而言可以比較各學區的平均就學距離的社會代價指數與跨越大型屏障的危險指數等評比原則。
Region planning has been an important issue around our societal activities. Before the invention of computer, we solved this tedious problem manually. Nowadays, the computer-aided region planning techniques have been applied widely on many planning practices, such as school district, convenience store location, range of police patrol, and irrigation area of farmland and public facilities. In the past, some solutions proposed might provide “acceptable” results, for example, the popular Voronoi method. But those techniques only focused on generating the location information of the regional center, and failed to include the spatial constraints, such as street width, river dimension or railroad barriers. Therefore, they may not suitable for some applications that require strong consideration of the cost of travel and safety problems. Suppose, we have a young kid dwelled along a street of 100 meters wide, with school A just across the street and school B about 500 meters far but in the same side of the road, should we assign the kid to school A or school B? Normally the kid was assigned to school A initially by a computer-aided software, since the travel distance is shorter. Latter with safety problem, the kid may reassign by hand to school B by the request of the parents. During our research process, we proposed a new “Terrace algorithm” that automatically counts the location of the regional center and its surround environment, so that it is more appropriate for solving similar problems as described on the above situation. Besides, we developed some criteria in comparing the suitability of some proposed techniques. In case of school district planning, the comparison is based on the cost and dangerous indexes of crossing over huge blockades. |
