**2.1. Methodology**

**1. Introduction**

318 Sustainable Urbanization

Rapidly increasing population induces growing cities and increasing car ownership. Conse‐ quently, transportation and land use problems become significant issues due to their econom‐ ical effects. Wey and Hsu [1] stated that urban sprawl and city congestion have become the inevitable development trend in the process of economic growth. Transportation and land use problems are getting inextricable issues simultaneously with the existing development trend.

Many researchers found that density is a significant factor of conversation of energy, and many studies found that access to high-capacity transit, incentives for development, balanced parking policy, mixed-use designs and jobs-housing balance are critical parameters of sustainability [2–7]. Transport-oriented problems and land use planning problems are directly

Interaction between land use and transportation is the basic factor for the trip generation. Transportation investments still strongly affect land use patterns, urban densities and housing prices [8]. Transportation systems primarily support sprawl [9]. To set up an effective transportation system, land use decisions have to be taken effectively and residential area densities have to be well arranged. Handy [10] stated that building more highways will contribute to more sprawl and lead to more driving. Building up a solitary transit-oriented system is not an exact solution. Conventional planning paradigm primarily builds the environment and afterwards tries to overcome the existing transportation problems [11]. Effects of land use decisions which generate strong travel attractiveness should be measured in the planning process since land use decisions acquire an irreversible characteristic after construction period. The interaction between land use and transportation should be meas‐

Effects of the increasing traffic volumes should be investigated by traffic impact analyses in order to find out whether the existing link capacities are convenient or not. Conventional land use planning paradigm is inclined to generate land use decisions by evaluating social and economical parameters [12]. Whereas, traffic impact analyses should be evaluated as one of the basic elements of land use planning parameter set. Those types of deficiencies bring along new planning paradigms such as new urbanism and smart growth (SG). The rise of new urbanism brings new energy and ideas to communities that commit to manage growth. Urban design hence becomes more visible within planning since the design is incorporated into growth management programs. Comprehensive planning also begin to connect more strongly with affordable housing advocates and public health professionals, broadening their focus beyond the more traditional set of issues revolving around land-use, transportation and the environment [13]. New urbanism is synonymous with SG, but there are significant differen‐ ces. New urbanism was much more influenced by architects and physical planners, while the SG was launched from a community of environmentalists, citizen groups, transportation

Many urban problems have led to a more intelligent and sophisticated planning trend which directly effects urban sprawl. Those problems may be stated as air and water pollution, loss

interdependent fields and have a highly interrelated iterative interaction.

ured and managed through traffic impact analyses.

planners and policy makers [14].

The main purpose of this study is to apply SG strategies to the land use planning process and evaluate the accuracy of land use planning decisions in the perspective of sustainable transportation. In this context a four-step land use planning procedure is proposed. The flow chart of the proposed procedure is given in **Figure 1**.

As can be seen in **Figure 1** that the proposed procedure starts with the initialization of the problem parameters. Land use pattern of the study area, travel demand between all origin– destination (O-D) pairs, transportation network characteristics such as link capacities, free flow travel times and signal timings for signalized intersections are provided in Step 1.

**Figure 1.** Four-step procedure.

In Step 2, Scenario-I that represents the conventional land use planning paradigm is ana‐ lyzed. In this context, a traffic assignment is carried out in order to calculate the link traffic volumes. In the developed land use planning procedure, stochastic user equilibrium (SUE) traffic assignment is proposed since drivers' perception errors are taken into account while they make their route choice decisions.

Considering a road network with sets of nodes *N*, directed links *A*, O-D pairs *W*, routes *P*, the SUE link traffic volumes may be calculated by solving Eq. (1) [20].

$$\underset{\mathbf{v}(\boldsymbol{\Psi})}{\text{Minimise}}\quad Z(\mathbf{v}(\boldsymbol{\Psi}),\boldsymbol{\Psi}) = -\mathbf{q}^{\mathrm{T}}\mathbf{y}(\mathbf{v}(\boldsymbol{\Psi}),\boldsymbol{\Psi}) + \mathbf{v}^{\mathrm{T}}\mathbf{t}(\mathbf{v}(\boldsymbol{\Psi}),\boldsymbol{\Psi}) - \sum\_{a \in \mathcal{A}} \int\_{0}^{v\_{a}(\boldsymbol{\Psi})} \mathbf{f}\_{a}(\boldsymbol{\Psi},\boldsymbol{x})d\mathbf{x} \tag{1}$$

subject to

$$\mathbf{h} \cdot \mathbf{q} = \mathbf{A} \mathbf{h}, \qquad \mathbf{v} \left( \mathbf{t} \boldsymbol{\psi} \right) = \boldsymbol{\delta} \mathbf{h}, \qquad \mathbf{h} \ge \mathbf{0} \tag{2}$$

where **q** is the vector of O-D demands [*qw*; ∀ *w* ∈ *W*], **v**(**ψ**) represents the vector of link traffic volumes, **ψ**is the vector of signal timings, **h** is the vector of route traffic volumes [*hp*; ∀ *p* ∈ *P*], *hp* is the traffic volume on route *p*, **y**(**v**(**ψ**), **ψ**) represents the vector that consists of travel times on all routes [*yp*; ∀ *p* ∈ *P*], **t**(**v**(**ψ**), **ψ**) is the vector of link travel times, [*yp*; ∀ *p* ∈ *P*] is the travel time along link *a*, *va* is the flow on link *a*, while **Λ** is the O-D/route incidence matrix [*Λp*; ∀ *p* ∈ *P*] and **δ**represents the link/route incidence matrix where *δap* =1 if link *a* is on route *p* and *δap* =0 otherwise [*δap*; ∀ *a* ∈ *A*; ∀ *p* ∈ *P*].

Eq. (1) can be solved by the path flow estimator (PFE) which is a traffic assignment tool using logit route choice model [21–25]. The solution procedure of PFE is given in **Figure 2**.

**Figure 2.** Flowchart of the PFE.

**Figure 1.** Four-step procedure.

320 Sustainable Urbanization

**ψ**

subject to

they make their route choice decisions.

In Step 2, Scenario-I that represents the conventional land use planning paradigm is ana‐ lyzed. In this context, a traffic assignment is carried out in order to calculate the link traffic volumes. In the developed land use planning procedure, stochastic user equilibrium (SUE) traffic assignment is proposed since drivers' perception errors are taken into account while

Considering a road network with sets of nodes *N*, directed links *A*, O-D pairs *W*, routes *P*, the

<sup>0</sup> ( ) ( ,) *<sup>a</sup> <sup>v</sup>*

where **q** is the vector of O-D demands [*qw*; ∀ *w* ∈ *W*], **v**(**ψ**) represents the vector of link traffic volumes, **ψ**is the vector of signal timings, **h** is the vector of route traffic volumes [*hp*; ∀ *p* ∈ *P*], *hp* is the traffic volume on route *p*, **y**(**v**(**ψ**), **ψ**) represents the vector that consists of travel times on all routes [*yp*; ∀ *p* ∈ *P*], **t**(**v**(**ψ**), **ψ**) is the vector of link travel times, [*yp*; ∀ *p* ∈ *P*] is the travel

*Z t x dx*

**ψψ ψψ ψψ ψ** *,* <sup>y</sup>

( )

*a A*

*a*

(1)

(2)

SUE link traffic volumes may be calculated by solving Eq. (1) [20].

( ) ( ) () () T T *Minimise* ( )( ) <sup>Î</sup> <sup>=</sup> -+- åò **<sup>v</sup> v qyv , vtv ,**

As can be seen in **Figure 2**, new route flows are calculated based on the logit route choice model. In this model, α is the dispersion parameter which controls the sensitivity of the route choice to the route travel times. Note that the convergence criterion *κ* that is based on flow similarity is used as given in Eq. (3) [26].

$$\frac{\sqrt{\sum\_{a} \left(\upsilon\_{a}^{u+1} - \upsilon\_{a}^{u}\right)}}{\sum\_{a} \upsilon\_{a}^{u}}\tag{3}$$

In applications, the value of the convergence criterion for the PFE solution may be accepted as 0.01 [27, 28]. After obtaining the link traffic volumes, network performance indicators are calculated for base-case and projection year under Scenario-I. In this study, VISSIM traffic simulation software is used for both visual analyses of the traffic and quantitative evalua‐ tion of the performance indicators which are average delay time per vehicle (seconds), average speed (km/h), average number of stops per vehicles, average stopped delay per vehicle (seconds), total delay time (hours), number of stops, total stopped delay (hours) and total travel time (hours).

In Step 3, new land use decisions are taken based on SG strategies under Scenario-II. Then residential area densities are modified by considering the land use plan of the city. At the evaluation process, economical, social, spatial and cultural factors can be considered. Afterwards, O-Ddemand matrix isupdateddirectlyproportionalto the new landusedecisions and then a SUE assignment is carried out to calculate link traffic volumes for projection year. As it was done in Step 2, the traffic is simulated on the road network and the performance indicators are calculated for Scenario-II.

In Step 4, Scenario-I and Scenario-II are compared in terms of the network performance indicators, and the new land use decisions are evaluated.

#### **2.2. Study area**

Denizli is an industrial metropolitan city which is located at the Aegean Region of Turkey with a population of over 600,000 in central district. It is also a tourism city and consists of 80 traffic analysis zones which were the administrative neighborhood districts before new governmen‐ tal regulations. The transport demand consists of mixed traffic which is supplied by private car, bus, minibus, service vehicle and taxi modes. Traffic problems increase in recent years in Denizli due to the high density of private car use [19]. The car ownership rate is about 22% which is about two times higher than the average car ownership in Turkey. The peak hour trips (07:00–09:00 a.m.) represent about 30% of the total trips which has been obtained by household surveys. The traffic analysis zones of the city are given in **Figure 3**.

**Figure 3.** Zonal layout.

**Figure 3** shows the zonal structure ofthe city. Inherently, land use densities are relatively lower and the zone sizes are much larger at the outer boundaries of the city. The major traffic problems are intersection delays. Therefore, a main signalized intersection serving heavy traffic volumes between three major arterials has been selected as the field of study. The aerial pictures of the selected intersection are given in **Figure 4**.

**Figure 4.** Illustration of the study intersection **(a)** and queue occurrence **(b)**.

As can be seen from **Figure 4a** that the study intersection is a signalized roundabout with four entry lanes on each approach. **Figure 4b** shows the queue occurrence on an approach with three isolated lanes that join the downstream link right after the roundabout. It is obvious that the performance of the intersection will decrease and lead very high level of traffic conges‐ tion considering the increase in future travel demand.
