Appunti di lezione Transport Planning

TRANSPORT PLANNING

The focus of the planning activity is on the question "should we do?"

• MANAGERIAL PERSPECTIVE - The real challenge is to take the right decisions in the transportation sector.
• The transport system is composed by two main elements:
  • The travel demand = need of people performing different activities in different places
  • The transport supply = infrastructures and transport services that can satisfy the travel demand

Engineers need to apply economics principles, according to the micro-economics.

FLOWS represent the interaction between travel demand and transportation supply ➔ transport planning has to deal with CONGESTION

In fact, transportation systems have a certain capacity. After this value of flow, the price increases without the increasing of quantities (flows).

• Congestion problems mainly concern urban areas, and worsened in 20s and 30s of XX century with the expansion of cars, as mean of transport affordable by the majority of the middle class ➔ Henry Ford in the USA. In Europe, regimes promoted the common use of cars and, at the same time, the construction of infrastructures.
• [supply-side policy, but without any planning] + [economic boom in the '50] ➔ mass motorization and congestion

1) In the USA during the '50s as many cars were stuck in the traffic, a lot of multi-lane roads were built but this solution did not solve the congestion, but worsened the situation: both the demand and supply were moved to the right.

• Necessity of transport planning, according to the principle predict and provide (= predict flows and consequently provide the infrastructures), in opposition to the "supply-side reactive policy" → Forecasting, substitutes reacting

In Detroit and Chicago, the first transport plans were made during the '50s, sponsored by public administrations and private car producers (e.g. General Motors)

2) In the '70s economic crisis due to the oil shock. Also the transportation sector had to change paradigm, adopting the so-called demand-side planning policy (= action on the demand curve).

1. Limitation of the travel demand
2. Multi-modal perspective substitutes car-centered vision

Need to contain the consumption of energy (oil crisis): there is the need to redirect the demand to more energy efficient means of transport.

3) In the '90s environmental concerns. An even stronger focus is put on alternatives to cars → Travel demand management strategies.

For these reasons: transport planning necessity.

iii) Multi-protocol survey

mixes the two protocols and tries to catch the best from each one of them.

• Regarding the survey periodicity
• Cross sectional is carried out in specific moments. It’s a picture in a certain time window.
• Continuous goes on for a long period / forever
• Panels carried out on several waves in which the same individuals answer within every wave. It allows to identify the dynamics of the mobility behavior in time.

QUESTIONNAIRE CONTENTS

1. TRAVEL DIARY

   collect all the movements within a certain period. ^{24h window for short trips / 1 month window for long trips}

   • List all trips in terms of origin, destination, purpose^{activity at the trip end}, time, travel means, cost^{out of pocket}, distance

   The trip could be split in stages. What is the mean that requests more loss of time?

   This info allows to better understand the priority and the purpose of a trip.

2. SOCIO-ECONOMIC CHARACTERISTICS OF INTERVIEWEE

   sex, age, education

   In order to calibrate the simulation model and evaluate future scenarios: households structure, built environment, car availability...

3. OPINIONS AND STATED PREFERENCES ON MOBILITY

   ask people their stated preferences

   Means to provide hypothetical scenarios ^(S.P.) in opposition to revealed preferences ^(R.P.).

NB: CENSUS never done on travel survey (impossible to involve all the people and observe all the trips). Not worth it.

• vs SAMPLE SURVEY population from which extract the sample (statistical operation)

Transport planners usually stop at this level. Otherwise:

CAUSALITY ISSUE ► from the independent y to the dependent x variable. In social sciences, this causality relationship can be put in discussion. In statistical terms, we can say that correlation between phenomena DOES NOT implies causality.

NB: This issue has to be taken in to account in the analysis, for example, of the relationship between car ownership and new car-sharing services.

SPURIOUS CORRELATION ► “fake” correlation between phenomena that fit perfectly by a statistical point of view, even if their correlation does not make any sense.

MODEL MIS-SPECIFICATION ► don’t consider all the variables that impact on the outcome. It is the case when, for example, we consider an increasing relationship between trip rate and income but without take into account the urban context (center or suburbs).

MODEL CLASSIFICATION

• MOBILITY MODELS
  • Long-terms decision of the individuals, that have an impact on the daily travel patterns e.g.:
    • owning a car
    • Residential choice
    • Obtain a driving license
    • Public transport pass
    • Workplace
• TRAVEL MODELS
  • Travel demand (traffic flows):
    1. TRIP-BASED
    2. TOUR-BASED
    3. ACTIVITY-BASED

• AGGREGATE or DISAGGREGATE MODELS
  • Disaggregate = individuals or units (HH) considered separately.
  • Aggregate = mean values of individuals considered together (Aggregate models are preferable in case of predictive studies)

GROWTH FACTOR METHODS

• Need the P/A matrix input (observed values).
• We assume that the observed values in the cells t_ij vary, by generally growing according to a multiplicative factor.
  • UNCONSTRAINED ➜ uniform growth factor r.
  T_ij = r z · t_ij when ∑_i T_ij = r z · σ_i ≠ 0;
  • SIMPLY CONSTRAINED ➜ compute a set of growth factors for each row r.
    • r = O_i / σ_i ➜ T_ij = r z · t_ij when ∑_i T_ij = r z · σ_i = O_i.
    we have to keep into consideration the output (or part of it) of the generation model * For columns: δ = D_j / d_j
  • DOUBLY CONSTRAINED ➜ the growth factor is a combination of r and δ (rows and columns) ➜ average growth factor model, that would be unconstrained, must be adjusted ➜ not closed form
    • DOUBLY CONSTRAINED AVERAGE GROWTH FACTOR
    STEP 1. T⁽¹⁾_ij = r_i + δ_j t_ij / 2O_i⁽¹⁾ ∑_j T⁽¹⁾_ij ≠ O_iD_j⁽¹⁾ = ∑_i T⁽¹⁾_ij ≠ D_jSTEP 2. Adjustment factors a_i = O_i / O_i⁽¹⁾ b_j = D_j / D_j⁽¹⁾Re-compute the trips taking the average of the adjusted factors T⁽²⁾_ij = a_i + b_j T_ij / 2and O_i⁽²⁾ ∑_j T⁽²⁾_ij ≠ O_iD_j⁽²⁾ ∑_i T⁽²⁾_ij ≠ D_jSTEP n. We iterate the same procedure until we reach an acceptable approximation ➜ O_i ≅ O_i: that means:a_i ≅ 1 ∀ib_j ≅ 1 ∀j