İstanbul ulaşım planlaması çerçevesinde çekim modelinin tahmin güçlerinin incelenmesi ve bir yöntem önerisi
Başlık çevirisi mevcut değil.
- Tez No: 75573
- Danışmanlar: PROF. DR. HALUK GERÇEK
- Tez Türü: Yüksek Lisans
- Konular: İnşaat Mühendisliği, Civil Engineering
- Anahtar Kelimeler: Ulaşım planlaması, İstanbul, Transport planning, Istanbul
- Yıl: 1998
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Ulaştırma Ana Bilim Dalı
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: Belirtilmemiş.
Özet
ÖZET Ulaşım planlaması çalışmaları kapsamında oluşturulan ulaşım modelleri ile, gözönüne alınan bölgedeki ulaşım hareketleri, çeşitli matematiksel bağıntılar kullanılarak temsil edilmekte, başka bir deyişle modellenmektedir. Bu doğrultuda aşağıda aşamaları sıralanan dört aşamalı klasik ulaşım modelleri oluşturulmaktadır. Bu çalışmaya konu olan ve sözü edilen dört aşamalı modelin ikinci aşaması olan Yolculuk Dağıtım Modelleri'nin işlevi, bir önceki aşamada herbir trafik bölgesi için hesaplanan yolculuk yaratım ve çekim değerlerini kullanarak, bölgeler arası yolculukları belirlemek diğer bir deyişle yolculuk matrislerini oluşturmaktır.. Yolculuk yaratım/çekimi. Yolculuk dağıtımı. Türel ayrım. Trafiğin yollara atanması Günümüze kadar birçok dağıtım modeli süregelmekle beraber, bu çalışma kapsamında bugün en çok kabul gören dağıtım modeli olan Çekim Modeli ele alınmıştır. Aşağıda matematiksel ifadesi verilen modelin dayandığı temel şudur.“Herhangi iki trafik bölgesi arasındaki yolculuk miktarı (ty), bu iki bölge arasındaki süre mesafe ya da genelleştirilmiş maliyet olarak direnimin bir fonksiyonu (f(cij)) ile ters orantılı ve bu bölgelerin yaratım (gi) ve çekim (aj) güçleri ile doğru orantılıdır.”tjj = kj *kj *g; *aj *exp(/? *Cij) 1987'de Halcrow Fox ve 1996'da ise İstanbul Teknik Üniversitesi tarafından İstanbul genelinde yapılan model çalışmalarında oluşturulan çekim modelleri ile günümüzde bu modellere ilişkin olarak tartışılan ve aşağıda iki madde halinde sunulan sorunlar incelenmiştir. Ayrıca bu çalışma kapsamında bu modellere seçenek olabilecek, çekim modelinin farklı bir uygulaması da anlatılarak, İstanbul modelinde uygulanmıştır.. Çekim modeli kalibrasyonunda, kalibrasyon yılına ait mevcut durum ve koşulları yansıtan yolculuk uzunluk dağılımları kullanılmaktadır. Hedef yılına ait matrisler de, mevcut verilerle kalibre edilmiş bu modeller ile oluşturulmaktadır. Diğer bir deyişle bölgenin gelecekteki yolculuk uzunluk dağılımlarını etkileyebilecek düzeyde önemli bazı sosyoekonomik ve arazi kullanım yapısı ve ulaşım ağındaki gelişmeler modelin kendisine yansıtılamamaktadır.. Model kalibrasyonunda kullanılan bu yolculuk uzunluk dağılımları, gözönüne alınan tüm bölge için elde edilmiş ortalama değerlerdir. Ancak kentin kendi içindeki arazi kullanım yapısının farklılaşmasından dolayı, bazı bölgelerden yapılan yolculukların yolculuk uzunluk dağılımları arasında, model açısından önemli olabilecek düzeyde önemli farklar olmaktadır. Dolayısıyla tüm çalışma alanı için tek bir yolculuk uzunluk dağılımı kullanılarak kalibrasyon yapılması pek de gerçekçi olmamaktadır.
Özet (Çeviri)
SUMMARY As the first step of the convertional transport planning procedure, trip generation models are used to estimate the total number of trips emanating from a zone (generations) and those attracted to each zones (attractions). Generations and attractions provide an idea of the level of trip making in a study area, but this is often not enough for modelling and decision making. What is needed is a better idea of the pattern of trip making, origion and destination of trips, the modes of transport chosen and the routes taken. In this study the second step of this four-step transport modelling study, trip distribution modelling will be described in detail. A number of methods have been put forward over the years to distribute trips among destinations; some of the simplests which are called growth factor methods are only suitable for short-term tactical studies where no major changes in the accessibility provided by the network is envisaged. Because the basic assumption made in such models is that the distribution of horizon-year trips from a zone is proportional to the base year trip distribution pattern modified by the growth factors of the zones under consideration. Others which are called Gravity Models seem to respond better to changes in network cost and therefore suggested for longer term, strategic studies or for tactical ones involving important changes in relative transport prices. The gravity model has become a standart transportation planning procedure for estimating interzonal trip interchanges rather than the growth factor methods and in this study we will focus on the gravity model with some depth. The basic premise of the gravity model used in urban transport studies is that the amount of trips two zone i and j İs directly proprotional to the number of trips produced in zone i (gi), the number of trips attracted to zone j (aj), and inversely proportional to some function of the spatial seperation of the two zones This premise may be exspressed algebraically as fallows: Where; f(cij) :some function of spatial seperation (travel time, distance or generalized cost) of zones i and j and sometimes called the friction factor ki,kj balancing factors which ensure that sum of the rows wnd the columns of the matrix is equal to the number of generations and attractions t» oc p-. *a.* /(c) lj d>j J ij This expression may be rewritten in the fallowing way to express the gravity model ts=ki*kj*&*aj*^cs) Travel time is normally used as the measure of the spatial seperation of zones in the gravity model and the fallowing functional forms have been used to derive the travel xmtime magnitudes. İn this study only the first one of the functions given below is used as friction function During the calibration of the distribution model of the 1996 Istanbul Transportation Master Plan, a doubly constrained gravity model is used. The calibration procedure will be briefly explained below. The most widely used technique for calibrating the gravity model was developed by the Bureau of Public Roads. The first step of the calibration procedure involves the estimation of the zonal travel time for each zone. The network is coded in terms of centroids nodes and links separately for highway and transit networks. If the travel time on each link is known, than the minimum travel time paths between any centroid pair may be established. To find out the travel time on each link in the morning peak hour, as this master plan study is based on the morning peak hour travels, the observed matrix is assigned to highway network and the assigned speeds for each link is calculated for highway network. For the transit network the speed for each link is multiplied by 0.80 to reflect the stops at the stations for highway public vehicles but for rails and ferries their exact speed is coded. From the household surveys made in 1996, it is known that the study area model split is 60% public transport and 40% private transport. Then the minimum travel time matrix which is called“skimtrees”between the zones used in this process is calculated by the weighted average given with the equation below, Qj =0.60*(public transport)+0.40(private transport) The criterion used to identify the correct travel time factor relationship is the trip length distribution. With an assumed initial travel time factor relationship is used to calculate trip interchange matrix and this set of trip interchanges may be used along with the skim trees to derive a trip length frequency distribution. The key to the calibration of the gravity model is to vary the travel time factor until the trip length frequency distribution simulated by the gravity model approximates that actually observed for city being studied. It is suggested that the gravity model simulated and observed trip length frequency distributions should exhibit the fallowing characteristic, the difference between two percent of trips for relative time range should be between ±3 percent. If the trip length frequency distribution produced by the gravity model does not fit this criterion, then a new set of travel time factors may be estimated from the fallowing expression f(cij) = f(cij)'*OD%/GM% XIVf(c<j) :the travel time factor for a given tavel time to be used in the next iteration f(cij)' :the travel time factor used in the calibration just completed OD% :the percentage of total trips occuring for a given travel time obsedved in the travel survey GM% : the percentage of total trips occuring for a given travel time simulated by the gravity model The trip matrix simulated by a gravity model which has been calibrated in the manner just described will not necessarily satisfy generation and attraction constraints given below which indicate that row and column totals of a trip matrix should be equal to the number of trips generated by the origin zone and the number of trips attracted by the destination zones. By using these formulas in an iterative approach, ki and kj balancing factors are calculated. N Ztjj = G= Trip generation constraint j Et|j = Aj Trip attraction constraint Where Gi and Aj are the future trip generations and attractions of zones Based on the experience of previous studies, it has been found out that travel time functions f(cjj) differ for different trip purposes. Consequently, the calibration of trip distribution is done for each trip purpose. In the 1996 Istanbul Master Plan calibration is carried out for the trip purposes given below.. Home based work trips. Home based school trips. Home based other trips. Non home based trips For each of four trip purposes, trip length frequency distributions and mean trip time of observed and simulated by the gravity model are calculated and compared in 10 minutes time ranges. From this comparesion it is observed that the fit achieved by the distribution model in terms of the trip length distributions mean trip time and also when synthesised matrices are compared with the observed ones they are close enough to each other within the allowed percentage of error. In 1987 a similar study was carried out for the same area İstanbul-city. In this study gravity model had been calibrated fallowing the same procedure which was explained for 1996-year calibration. The only difference was that friction function was a function of genarelized cost not travel time like 1996 year gravity model. As one of the objectives of this study is to compare the models calibrated in 1987 and 1996, they should be compatible with each other. Therefor 1987-year gravity model was calibrated by travel time in the same way, which was explained for 1996-year gravity model. These different master plan studies, that have been made for the same area istanbul- city provide a basis for comparison of the results of the both distribution models XVcalibrated in different years. First, the observed trip length frequency distribution and mean trip time were compared to find out how trip patterns and travel characteristics changed during this period of time. By using the results of household surveys made in 1987 and 1996 it can be seen that the mean trip time from 1987 to 1996 have been decreased for each four trip purposes. And also the percentage of the number of trips with shorter travel time has been increased. As a result of this decrease in mean trip time the friction function and the calibration constant“P”have also decreased. We can briefly say that in the past nine years in Istanbul with the increasing worsening conditions of traffic, people preferred to live close to where they work or where they go to school or the number of employment opportunities began to spread out from the centre of the city to suburban areas which are residentially densitied areas. This is an expected result for such a big and density city with 9 millions population. But the important issue is to figure out how such changes will affect the results of the model calibration results and the performance of the model. As an objective of this study to show the performance the gravity models, a study has been made by using both models which will be explained below. Although gravity model is most widely used model for trip distribution modelling, there are some important problems. Some of these problems which have been dealed with in this study can be summarised in two groups.. The BPR procedure assumes that friction function will remain constant over the time in the future. As the pattern of the urban development in the city or region may change considerably in this time period, it is questionable if the form of the function will remain constant.. Another principal difficulty with the calibration procedure is that travel friction function and the associated trip length frequency distributions are assumed to be constant for each zone of a study area. It is clear to see that this assumption is not valid especially in the big study areas like Istanbul significant variations occur troughout the whole area. In some similar studies row specific travel time functions have been calibrated, in another words,“P”calibration constant have been calculated. Another way of achieving this would be to give suburban-suburban trips one trip length distribution frequency and central business district to suburban district another trip length distribution frequency and to calibrate the curves separately. But the problem mentioned above outstands here again as, with the changing of base year conditions in the future, some suburban part of the city may turn to a central business area. In this study, to test the performance of the gravity models a study has been made by using both models calibrated in 1987 and 1996. This also gives us a chance to see and test the results of the models that are used to estimate the travel demand in the long term. To test the performance of the 1987-year gravity model, 1996-year trip demand was estimated by using this model. 1996-year skimtree between the zones was calculated by using the current network. Taking the 1996-year skimtree and observed trip generation and attractions of zones as input data for the 1987-year gravity model, 1996-year matrices were calculated. These model matrices were compared with the XVI1996-year model calibration results; it can be seen that 1987 model results are not worse than the 1996-year results. Although there have been some changes in trip length distributions and“0”friction factor constants when compared with the 1996 year results But the point here is that 1996 network and generation and attractions were used. Therefore no errors outstands from not being able to propose correctly the future developments. As a result of this, it can be said that in such strategic planning and modelling studies the errors arising from wrong proposal of the future developments are much bigger than the errors which stand from the model itself. To deal with the problem which was mentioned secondly an alternative study to 1996-year gravity model has been made as one of the objective of this study. A multisectorial gravity model with five sectors with row specific calibration constants is calibrated for the same study area Istanbul. During this procedure the most important issue is to decide which zones will be included in each of the five sectors. For the grouping of zones, an expression called Total Average Weighted Time (TAS) which is based on the expression of accessibility was used. Total average weighted time of zone i (T AS;) is the ratio of the sum of the weighted average of the travel time between two zones (Cy) with the number of the opportunities as number of employment or number of students (Pi) according to the trip purpose to the total number of opportunities over that area (P) This expression may be given algebraically as fallows, N Zc*P. TAS; = J-- It can easily be seen from this expression that TAS reflects the distribution of employment or school densitied zones which shapes the curve of trip length frequency distributions. Therefor it makes sense to use this formulation to group and separate zones into gravity model sectors. From the point that the zones with less TAS is closer to the highly opportunity densitied, five ranges of TAS were described and the zones of which their TAS is involved in the range described will be included in that sector. From the calibration results which have been made seperately for each sector it is seen that the sectors with less TAS have less mean trip time and less number of long trips when compared with the sectors with higher TAS. As result of this, such sectors with different TAS would have different value of calibration constant“P”.And also the distribution of the zones with less TAS show the same pattern with highly employment densitied zones. From all these results it can easiliy be said that using multi sectorial gravity model based on the TAS is more realistic and would reflect the trip patterns better than the other. xvnBut the most important result drawn from this study is that any changes in future conditions that may be important for the model are reflected with changing TAS values. Depending on the amount of the change in TAS value of a zone, that zone may be included in another more appropriate sector as a result of the pattern of the development of the area undertaken. In another words model will be more sensible to the developments occurred in the study area. xvm
Benzer Tezler
- Metropoliten alanlarda kentiçi ulaşım yönetimi İstanbul örneği
The Innercity transportation management in metropolitan areas İstanbul as a sample
PERVİN BETÜL ZENGİN AYTAÇ
Yüksek Lisans
Türkçe
1997
Şehircilik ve Bölge Planlamaİstanbul Teknik ÜniversitesiŞehir ve Bölge Planlama Ana Bilim Dalı
PROF. DR. YÜCEL ÜNAL
- Kent ve peyzaj planlaması çerçevesinde kıyı dolgu düzenlemelerinin incelenmesi; Maltepe-Bostancı ve Samatya dolgu alanları sahil parklarının peyzaj planlama açısından irdelenmesi
Investigating the arrangement coast means of urban and landscope planning: Maltepe-Bostancı and Samatya filled coastal parks taken as an example for the landscape investigation
KEVSER ÖKSÜZ
Yüksek Lisans
Türkçe
2000
Şehircilik ve Bölge Planlamaİstanbul Teknik ÜniversitesiPROF.DR. AYŞE SENA KUBAT
- İstanbul Olimpiyat Oyunları organizasyon planlaması ve tesislerinin incelenmesi (2000-2004 adaylıkları çerçevesinde bir araştırma)
The Planning and research of facilities in association with the Istanbul tender for the 2000-2004 Olympic Games
SEMİH YILMAZ
Yüksek Lisans
Türkçe
1999
SporOndokuz Mayıs ÜniversitesiBeden Eğitimi ve Spor Ana Bilim Dalı
YRD. DOÇ. DR. M. YALÇIN TAŞMEKTEPLİGİL
- İskenderun Kenti Yeşil Alanlarının irdelenmesi
A Research of green areas in İskenderun
MİRNA ÖLMEZ
Yüksek Lisans
Türkçe
1995
Şehircilik ve Bölge Planlamaİstanbul Teknik ÜniversitesiPROF.DR. AHMET YILDIZCI
- A Configuration of systematic approaches for drinking water distribution problem in metropolitan areas
Başlık çevirisi yok
SELİM KAHVECİOĞLU
Doktora
İngilizce
1997
Mühendislik Bilimleriİstanbul Teknik Üniversitesiİnşaat Mühendisliği Ana Bilim Dalı
PROF. DR. SELİME SEZGİN