5.4 Transportation Planning, Network Analysis, and Intermodal Systems

5.4 Transportation Planning, Network Analysis, and Intermodal Systems

The AMSTHC TOS area "Transportation Engineering" has these competencies:

1. Review concepts in urban transportation planning.
2. Analyze networks.
3. Perform quantitative methods for transportation engineering.
4. Evaluate and assess intermodal transportation networks.
5. Understand methods used to prepare cost estimates. This one repeats a quantity-surveying line and is covered in Section 6.6.

Traffic-flow quantitative methods are in Section 5.2. This section covers planning and networks.

---

The Urban Transportation Planning Process

Transportation planning asks how many trips will be made, between which places, by which modes and on which routes. The classic four-step model answers these questions in sequence:

The study area is divided into traffic analysis zones. Travel is summarized in an origin-destination (O-D) matrix. Data come from household interview surveys, roadside O-D surveys, traffic counts and land-use forecasts.

---

Trip Generation

Productions ($P_i$) are trips generated by homes. Attractions ($A_j$) are trips drawn by jobs, schools and shops.

Example. A regression model is $P = 0.5 + 1.2(\text{household size}) + 0.8(\text{cars owned})$ daily trips per household. For 1,000 households averaging 4.5 persons and 0.5 car, zone productions are:

$1{,}000(0.5 + 5.4 + 0.4) = 6{,}300 \text{ trips/day}$

Total productions and attractions are then balanced so that the region's totals match.

---

Trip Distribution: The Gravity Model

$T_{ij} = P_i \frac{A_j F_{ij} K_{ij}}{\sum_j A_j F_{ij} K_{ij}}$

• $F_{ij}$ is a friction factor that decreases with travel time or cost, for example $F = t^{-b}$.
• $K_{ij}$ is an adjustment factor, often set to 1.

Example. Zone 1 produces 1,000 trips. Zone A attracts 600 trips at $t = 10\text{ min}$; zone B attracts 400 trips at $t = 20\text{ min}$. Use $F = t^{-2}$.

• Weights: $600/100 = 6.0$ and $400/400 = 1.0$.
• $T_{1A} = 1{,}000(6/7) = 857$ trips and $T_{1B} = 143$ trips.

Nearer, larger attractions get most of the trips.

---

Mode Choice: The Logit Model

Each mode $m$ has a utility $U_m$, usually a linear function of travel time, cost and comfort, with negative coefficients on time and cost. The multinomial logit share is:

$P_m = \frac{e^{U_m}}{\sum_k e^{U_k}}$

Example. $U_{\text{car}} = -0.05(30) - 0.02(80) = -3.10$ and $U_{\text{bus}} = -0.05(45) - 0.02(15) = -2.55$.

• $e^{-3.10} = 0.0450$ and $e^{-2.55} = 0.0781$.
• Bus share $= 0.0781/(0.0450 + 0.0781) = 0.634$, or 63%.

---

Network Analysis and Assignment

A transport network is a set of nodes (intersections, terminals) and links (road segments) with costs such as travel time.

Shortest path (Dijkstra's algorithm)

1. Give the origin a label of 0 and every other node a label of infinity.
2. Repeatedly make permanent the node with the smallest temporary label.
3. Update the labels of its neighbors: new label = permanent label + link time, if that is smaller.
4. Stop when the destination is permanent, and trace back the predecessor nodes.

Example. Links: A–B 4 min, A–C 2, C–B 1, B–D 5, C–D 8.

• From A: C = 2 is permanent; B = min(4, 2 + 1) = 3 is permanent.
• D = min(3 + 5, 2 + 8) = 8.
• The shortest path is A–C–B–D in 8 minutes.

Assignment methods

• All-or-nothing. All trips between two zones take the shortest path. This ignores congestion.
• User equilibrium (Wardrop's first principle). Each traveler picks the quickest route, so in equilibrium all used routes between an O-D pair have equal and minimum travel times. Link travel times rise with volume, for example through the BPR function $t = t_0[1 + 0.15(v/c)^4]$.
• System optimum (Wardrop's second principle). The total travel time of all users is minimized. It may require some users to take slower routes, so it is a benchmark for traffic management and pricing.

Equilibrium example. Two routes carry 1,000 veh/h, with $t_1 = 10 + 0.01x_1$ and $t_2 = 15 + 0.005x_2$ minutes. Setting $t_1 = t_2$ with $x_1 + x_2 = 1{,}000$:

$10 + 0.01x_1 = 15 + 0.005(1{,}000 - x_1) \implies 0.015x_1 = 10 \implies x_1 = 667$

So $x_2 = 333$, and both routes take $16.67\text{ min}$.

---

Intermodal Transportation Networks

Intermodal transport uses more than one mode for a single trip or shipment. Examples are jeepney to rail to walk, or truck to port to ship to truck. Total performance depends on the links and also on the transfer points.

Key measures:

• Door-to-door time includes access, waiting, in-vehicle and transfer times.
• Generalized cost combines money cost with time valued in pesos per hour.
• Reliability reflects variability of travel time.

Good intermodal design minimizes transfer penalties: short walking distances, coordinated schedules, integrated fares, and protected waiting areas.