Modelling Transport

Dr. Juan de Dios Ortúzar is Emeritus Professor in the School of Engineering at Pontificia Universidad Católica de Chile and also Key Researcher at Instituto Sistemas Complejos de Ingeniería (ISCI) and the BRT+ Centre of Excellence. He has over 30 years of experience in discrete choice modelling and survey design with particular focus on transport demand modelling and the valuation of transport externalities. Dr. Luis G. Willumsen is an internationally recognised authority in transport and traffic modelling and has over 30 years of experience in this area. He previously lectured at Leeds University and University College London, and was also a Director of Steer before leaving in 2009 to set up his own independent practice. He is also Managing Partner of Nommon Solutions and Technologies, a company processing big data to provide location and mobility intelligence.

• Preface xviiiAbout the Companion Website xxii1 Introduction 11.1 Background 11.2 Models and Their Role 21.3 Characteristics of Transport Problems 31.3.1 Characteristics of Transport Demand 31.3.2 Characteristics of Transport Supply 41.3.3 A View of Transport Problems 61.3.4 A Simple Model 71.3.5 Classic and New Modes of Transport 91.4 Modelling and Decision-Making 91.5 Issues in Transport Modelling 121.5.1 General Modelling Issues 121.5.1.1 The Roles of Theory and Data 121.5.1.2 Model Assumptions 131.5.1.3 Model Specification 141.5.1.4 Model Calibration, Validation, and Use 151.5.1.5 Modelling, Forecasting, and Judgement 161.5.2 Aggregate and Disaggregate Modelling 171.5.3 Homo Sapiens and Homo Economicus 181.5.4 Cross-Section and Time Series 201.5.5 Revealed and Stated Preferences 211.6 The Structure of the Classic Transport Model 221.6.1 The Classic 4/5 Stage Model 221.6.2 Granularity 241.6.3 Macro, Meso, and Micro Models 271.7 Transport Planning and Uncertainty 271.8 Theoretical Basis Versus Expedience 311.9 Becoming a Better Modeller 33Exercises 332 Data 352.1 Basic Sampling Theory 362.1.1 Statistical Considerations 362.1.1.1 Basic Definitions 362.1.1.2 Sample Size to Estimate Population Parameters 382.1.1.3 Obtaining the Sample 402.1.2 Practical Considerations in Sampling 432.1.2.1 The Implementation Problem 432.1.2.2 Finding the Size of Each Subpopulation 432.2 Errors in Modelling and Forecasting 442.2.1 Different Types of Error 452.2.1.1 Measurement Errors 452.2.1.2 Sampling Errors 462.2.1.3 Computational Errors 462.2.1.4 Specification Errors 462.2.1.5 Transfer Errors 472.2.1.6 Aggregation Errors 472.2.2 The Model Complexity/Data Accuracy Trade-off 482.2.3 Forecasting Errors 512.3 Basic Data-collection Methods 532.3.1 Practical Considerations 532.3.1.1 Length of the Study 532.3.1.2 Study Horizon 532.3.1.3 Limits of the Study Area 542.3.1.4 Study Resources 542.3.2 Types of Surveys 542.3.2.1 Survey Scope 552.3.2.2 Home Interview Travel Surveys 572.3.2.3 Other Important Types of Surveys 662.3.3 Data Correction, Expansion, and Validation 682.3.3.1 Data Correction 692.3.3.2 Imputation Methods 712.3.3.3 Sample Expansion 722.3.3.4 Validation of Results 722.3.4 Longitudinal Data Collection 732.3.4.1 Basic Definitions 732.3.4.2 Representative Sampling 742.3.4.3 Sources of Error in Panel Data 752.3.4.4 Relative Costs of Longitudinal Surveys 762.3.5 Travel Time Surveys 762.3.6 Digital Data Sources 772.4 Stated Preference Surveys 792.4.1 Introduction 792.4.1.1 Contingent Valuation and Conjoint Analysis 792.4.1.2 Stated Choice Methods 812.4.2 The Survey Process 832.4.2.1 Clarifying Study Objectives and Defining Objects of Interest 842.4.2.2 Defining Experimental Assumptions 862.4.2.3 Generating the Experimental Design 922.4.2.4 Conduct Post Design Generation Testing 972.4.2.5 Conduct Questionnaire 982.4.2.6 Nothing is Important 992.4.2.7 Realism and Complexity 1002.4.2.8 Use of Computers in SP Surveys 1012.4.2.9 Quality Issues in Stated Preference Surveys 1022.4.3 Case Study Example 1032.4.4 Limitations of Stated Preference Methods 115Exercises 1153 Zones and Networks 1193.1 Zoning Design 1203.2 Road Network Representation 1223.2.1 Traffic Flow 1233.2.2 Network Details 1233.3 Link Properties and Functions 1253.3.1 Link Properties 1253.3.2 Network Costs 1263.3.3 Definitions and Notation 1273.3.4 Speed-Flow and Cost-Flow Curves 1273.3.5 Public Transport Networks 131Exercises 1324 Trip Generation Modelling 1334.1 Introduction 1344.1.1 Some Basic Definitions 1344.1.2 Characterisation of Journeys 1354.1.2.1 By Purpose 1354.1.2.2 By Time of Day 1354.1.2.3 By Person Type 1364.1.3 Factors Affecting Trip Generation 1364.1.3.1 Personal Trip Productions 1374.1.3.2 Personal Trip Attractions 1374.1.3.3 Freight Trip Productions and Attractions 1374.1.4 Growth-Factor Modelling 1384.2 Regression Analysis 1394.2.1 The Linear Regression Model 1394.2.2 Zonal-Based Multiple Regression 1484.2.3 Household-Based Regression 1494.2.4 The Problem of Non-Linearity 1514.2.5 Obtaining Zonal Totals 1524.2.6 Matching Generations and Attractions 1534.3 Cross-Classification or Category Analysis 1534.3.1 The Classical Model 1534.3.1.1 Introduction 1534.3.1.2 Variable Definition and Model Specification 1544.3.1.3 Model Application at Aggregate Level 1554.3.2 Improvements to the Basic Model 1564.3.2.1 Equivalence Between Category Analysis and Linear Regression 1564.3.2.2 Regression Analysis for Household Strata 1584.4 Other Trip Generation Formulations 1594.4.1 Alternative Model Formulations 1594.4.1.1 The Negative Binomial (NB) Approach 1594.4.1.2 The Ordinal Probit Model 1604.4.1.3 Comparing the Performance of Count Data and Linear Regression Models 1604.5 Trip Generation and Accessibility 1614.6 The Frequency Choice Logit Model 1624.7 Tour Generation 1644.8 Forecasting Variables in Trip Generation Analysis 1654.9 Stability and Updating of Trip Generation Parameters 1674.9.1 Temporal Stability 1674.9.2 Geographic Stability 1684.9.3 Bayesian Updating of Trip Generation Parameters 168Exercises 1715 Trip Distribution Modelling 1735.1 Definitions and Notation 1745.2 Growth-Factor Methods 1765.2.1 Uniform Growth Factor 1765.2.2 Singly Constrained Growth-Factor Methods 1775.2.3 Doubly Constrained Growth Factors 1785.2.4 Advantages and Limitations of Growth-Factor Methods 1805.3 Synthetic or Gravity Models 1805.3.1 The Gravity Distribution Model 1805.3.2 Singly and Doubly Constrained Models 1825.4 The Entropy-Maximising Approach 1835.4.1 Entropy and Model Generation 1835.4.2 Generation of the Gravity Model 1855.4.3 Properties of the Gravity Model 1875.4.4 Production–Attraction Format 1895.4.5 Segmentation 1905.5 Calibration of Gravity Models 1905.5.1 Calibration and Validation 1905.5.2 Calibration Techniques 1915.6 The Tri-Proportional Approach 1925.6.1 Bi-Proportional Fitting 1925.6.2 A Tri-Proportional Problem 1945.6.3 Partial Matrix Techniques 1955.7 Other Synthetic Models 1975.7.1 Generalisations of the Gravity Model 1975.7.2 Intervening Opportunities Model 1985.7.3 Disaggregate Approaches 2005.8 Practical Considerations 2005.8.1 Sparse Matrices 2005.8.2 Treatment of External Zones 2015.8.3 Special Generators 2015.8.4 Intra-Zonal Trips 2015.8.5 Journey Purposes 2025.8.6 K Factors 2025.8.7 Adjusting Trip Matrices 2035.8.8 Errors in Modelling 2035.8.9 The Stability of Trip Matrices 2045.8.10 Sense Checks 206Exercises 2066 Modal Split and Direct Demand Models 2096.1 Introduction 2096.2 Factors Influencing the Choice of Mode 2096.3 Trip-End Modal-Split Models 2116.4 Trip Interchange Modal-Split Models 2116.5 Synthetic Models 2136.5.1 Distribution and Modal-Split Models 2136.5.2 Distribution and Modal-Split Structures 2156.5.3 Multimodal-Split Models 2166.5.4 Calibration of Binary Logit Models 2196.5.5 Calibration of Hierarchical Modal-Split Models 2206.6 Direct Demand Models 2226.6.1 Introduction 2226.6.2 Direct Demand Models 2226.6.3 An Improvement on Direct Demand Modelling 2246.7 Sense Checks 225Exercises 2277 Discrete Choice Models 2317.1 General Considerations 2317.2 Theoretical Framework 2347.3 The Multinomial Logit (MNL) Model 2367.3.1 Specification Searches 2387.3.2 Universal Choice Set Specification 2397.3.3 Some Properties of the MNL 2407.4 The Nested Logit Model (NL) 2417.4.1 Correlation and Model Structure 2417.4.2 Fundamentals of Nested Logit Modelling 2427.4.2.1 The Model of Williams and of Daly–Zachary 2437.4.2.2 The Formulation of McFadden: The GEV Family 2447.4.3 The NL in Practice 2467.4.3.1 Limitations of the NL 2477.4.4 Controversies About Some Properties of the NL Model 2477.4.4.1 Specifications Which Address the Non-Identifiability Problem 2477.4.4.2 On the Limits of the Structural Parameters 2497.4.4.3 Two Further Issues 2517.5 The Multinomial Probit Model 2537.5.1 The Binary Probit Model 2537.5.2 Multinomial Probit and Taste Variations 2547.5.3 Comparing Independent Probit and Logit Models 2557.6 The Mixed Logit Model 2557.6.1 Model Formulation 2557.6.2 Model Specifications 2567.6.2.1 Basic Formulations 2567.6.2.2 More Advanced Formulations 2587.6.3 Identification Problems 2597.6.3.1 Theoretical Identification 2607.6.3.2 Empirical Identification 2607.7 Other Choice Models and Paradigms 2617.7.1 Other Choice Models 2617.7.2 Choice by Elimination and Satisfaction 2627.7.2.1 Compensatory Rule 2637.7.2.2 Non-Compensatory Rules 2637.7.3 Habit and Hysteresis 2647.7.4 Modelling with Panel Data 2657.7.4.1 Panel Data Models 2667.7.4.2 Efficiency and Repeated Observations 2677.7.4.3 Dealing with Temporal Effects 2697.7.5 Hybrid Choice Models Incorporating Latent Variables 2717.7.5.1 Modelling with Latent Variables 2727.7.5.2 Hybrid Discrete Choice Model 2727.7.6 Attribute Non-Attendance and Other Heuristics 273Exercises 2768 Specification and Estimation of Discrete Choice Models 2798.1 Introduction 2798.2 Choice-Set Determination 2808.2.1 Choice-Set Size 2808.2.2 Choice-Set Formation 2818.3 Specification and Functional Form 2828.3.1 Functional Form and Transformations 2828.3.1.1 Basic Box–Cox Transformation 2838.3.1.2 Box–Tukey Transformation 2838.3.2 Theoretical Considerations and Functional Form 2838.3.3 Intrinsic Non-Linearities: Destination Choice 2848.4 Statistical Estimation 2858.4.1 Estimation of Models from Random Samples 2858.4.1.1 The t-test for Significance of any Component θ∗k of θ∗ 2888.4.1.2 The Likelihood Ratio Test 2918.4.1.3 The Overall Test of Fit 2928.4.1.4 The ρ2 Index 2938.4.1.5 The Percentage Right or First Preference Recovery (FPR) Measure 2948.4.1.6 Working with Validation Samples 2958.4.2 Estimation of Models from Choice-based Samples 2998.4.3 Estimation of Hybrid Choice Models with Latent Variables 3008.4.4 Comparison of Non-Nested Models 3048.4.5 Correcting for Endogeneity in Discrete Choice Models 3058.4.6 Accounting for Stochastic Variables in Choice Models 3078.4.6.1 Econometric Analysis 3098.4.6.2 Stochastic Variables Model 3108.4.6.3 Random Coefficients Model 3118.5 Estimating the Multinomial Probit Model 3118.5.1 Numerical Integration 3138.5.2 Simulated Maximum Likelihood 3148.5.2.1 The Basic Approach 3148.5.2.2 Advanced Techniques 3158.6 Estimating the Mixed Logit Model 3178.6.1 Classical Estimation 3178.6.1.1 Estimation of Population Parameters 3178.6.1.2 Estimating Individual Parameters 3188.6.2 Bayesian Estimation 3198.6.3 Choice of a Mixing Distribution 3238.6.3.1 Alternative Mixing Distributions 3248.6.3.2 Discrete Mixtures and Latent Class Modelling 3258.6.3.3 Empirical Identifiability of Latent Class Models 3278.6.4 Binary Choice Case 3278.6.5 Random and Quasi Random Numbers 3308.6.6 Estimation of Panel Data Models 3328.7 Modelling with Stated-Preference Data 3348.7.1 Identifying Functional Form 3348.7.2 Stated Preference Data and Discrete Choice Modelling 3368.7.2.1 Naive Methods 3378.7.2.2 Discrete Choice Modelling with Rating Data 3398.7.2.3 Discrete Choice Modelling with Rank Data 3398.7.2.4 Modelling with Stated Choice Data 3418.7.2.5 Model Estimation with Generalised Choice Data 3428.7.2.6 Modelling with Indifference Alternatives 3458.7.2.7 Interactions in SC Modelling 3518.7.2.8 The Problem of Repeated Observations 3548.7.3 Model Estimation with Mixed SC and RP Data 3558.7.3.1 Estimation without Considering Correlation among Repeated Observations 3558.7.3.2 Joint Estimation Considering Correlation between Repeated Observations 3588.7.3.3 Forecasting with Joint RP–SC Models 359Exercises 3629 Model Aggregation and Transferability 3659.1 Introduction 3659.2 Aggregation Bias and Forecasting 3669.3 Confidence Intervals for Predictions 3679.3.1 Linear Approximation 3689.3.2 Non-Linear Programming 3699.4 Aggregation Methods 3709.5 Model Updating or Transference 3739.5.1 Introduction 3739.5.2 Methods to Evaluate Model Transferability 3739.5.2.1 Test of Model Parameter for Equality 3749.5.2.2 Disaggregate Transferability Measures 3749.5.3 Updating with Disaggregate Data 3759.5.3.1 Updating the Constants 3769.5.3.2 Updating of Constants and Scale 3769.5.4 Updating with Aggregate Data 377Exercises 37810 Static Assignment 38110.1 Basic Concepts 38110.1.1 Introduction 38110.1.2 Traffic and Queues 38310.1.3 Factors Influencing Route Choice 38510.2 Static Traffic Assignment Methods 38610.2.1 Introduction 38610.2.2 Modelling Route Choice 38710.2.3 Tree Building 38810.3 All-or-Nothing Assignment 39010.4 Stochastic Methods 39210.4.1 Simulation-Based Methods 39210.4.2 Proportional Stochastic Methods 39310.4.3 Emerging Approaches 39510.5 Congested Assignment 39810.5.1 Wardrop’s Equilibrium 39810.5.2 Hard and Soft Speed-Change Methods 40010.5.3 Incremental Assignment 40110.5.4 Method of Successive Averages 40210.5.5 Braess’s Paradox 40310.6 Public-Transport Assignment 40510.6.1 Introduction 40510.6.2 Issues in Public-Transport Assignment 40510.6.2.1 Supply 40510.6.2.2 Passengers 40710.6.2.3 Monetary Costs 40710.6.2.4 The Definition of Generalised Costs 40710.6.2.5 The Common Lines Problem 40810.6.2.6 Frequency or Schedule-Based Route Choice 40810.6.3 Modelling Public-Transport Route Choice 40810.6.4 Assignment of Public Transport Trips 41210.6.5 Discrete Route Choice Modelling 41310.7 Limitations of the Classic Methods 41510.7.1 The Assumption of Perfect Information about Costs in all Parts of the Network 41510.7.2 The Assumption that all Movements can be Represented by a Trip Matrix 41510.7.3 Limitations in the Node-link Model of the Road Network 41510.7.4 Errors in Defining Average Perceived Costs 41610.7.5 Not all Trip Makers Perceive Costs in the Same Way 41610.7.6 Day-to-Day Variations in Demand 41710.7.7 Imperfect Estimation of Travel Time Changes with Link Flow Changes 41710.7.8 The Dynamic Nature of Traffic 41810.7.9 Input Errors 41810.8 Practical Considerations 418Exercises 42211 Dynamic Assignment 42511.1 Introduction 42511.2 Travel Time Reliability 42511.3 Junction Interaction Methods 42611.4 The Dynamic Nature of Traffic 42711.4.1 Delays over Time and Space 42711.4.2 Average and Experienced Travel Times 43111.5 Dynamic Traffic Assignment (DTA) 43211.5.1 General Requirements 43211.5.2 Discretising Time in DTA 43311.5.3 Micro- and Meso-Simulation 43311.5.4 Equilibrium and Simulation 435Exercises 43812 Equilibrium 43912.1 Introduction 43912.2 Equilibrium 43912.2.1 A Mathematical Programming Approach 44012.2.2 Social Equilibrium 44412.2.3 Solution Methods 44512.2.3.1 The Frank–Wolfe Algorithm 44612.2.3.2 Route-Based Assignment 44712.2.3.3 Origin-Based Assignment 44812.2.4 Stochastic Equilibrium Assignment 45012.2.5 Congested Public Transport Assignment 45112.3 Transport System Equilibrium 45212.3.1 Equilibrium and Feedback 45212.3.2 Formulation of the Combined Model System 45512.3.3 Solving General Combined Models 45812.3.4 Monitoring Convergence 459Exercises 46013 Departure Time Choice 46313.1 Introduction 46313.2 Macro and Micro Departure Time Choice 46313.3 Underlying Principles of Micro Departure TIME Choice 46413.4 Simple Supply/Demand Equilibrium Models 46613.5 Time of Travel Choice and Equilibrium Assignment 46713.6 Modelling Disaggregate Time of Day Choice 46913.7 Joint Mode/Time of Day Choice 47413.7.1 Data Collection 47413.7.1.1 Alternatives and Their Attributes 47513.7.1.2 Stated Preference Data 47513.7.2 Model Estimation 47613.7.2.1 Analysis of Results 47813.7.2.2 Model Valuations 47913.8 Conclusion 47914 Complementary Techniques 48114.1 Introduction 48114.2 Sketch Planning Methods 48214.3 Incremental Demand Models 48314.3.1 Incremental Elasticity Analysis 48414.3.2 Incremental or Pivot-Point Modelling 48514.4 Model Estimation From Traffic Counts 48814.4.1 Introduction 48814.4.2 Route Choice and Matrix Estimation 48914.4.3 Transport Model Estimation from Traffic Counts 48914.4.4 Matrix Adjustments Using Traffic Counts 49214.4.5 Traffic Counts and Matrix Estimation 49714.4.5.1 Independence 49714.4.5.2 Inconsistency 49814.4.6 Limitations of ME 2 49914.4.7 Improved Matrix Estimation Models 50114.4.8 Treatment of Non-Proportional Assignment 50214.4.9 Quality of Matrix Estimation Results 50314.4.10 Estimation of Trip Matrix and Mode Choice 50414.4.10.1 Simple Unimodal Case 50414.4.10.2 Updating with Aggregate Modal Shares 50514.4.10.3 Updating with Traffic Counts 50514.4.10.4 Updating with Combined Information 50514.5 Gaming Simulation 506Exercises 50815 Freight Demand Models 51115.1 A Subject of Increasing Importance 51115.2 Factors Affecting Goods Movements 51215.3 Pricing Freight Services 51315.4 Data Collection for Freight Studies 51415.5 Aggregate Freight Modelling 51515.5.1 Freight Generations and Attractions 51615.5.2 Distribution Models 51615.5.3 Mode Choice 51815.5.4 Assignment 51815.5.5 Equilibrium 51915.5.6 Freight and Service Trips 52015.6 Disaggregate Approaches 52115.7 Conclusions 52216 Activity-Based Models 52316.1 Introduction 52316.2 Activities, Tours, and Trips 52416.3 Tours, Individuals, and Representative Individuals 52716.4 Agent-Based Modelling 52816.5 Activity-Based Modelling 52916.5.1 Introduction 52916.5.2 Population Synthesis 53016.5.3 Monte Carlo and Probabilistic Processes 53316.5.4 Structuring, Activities, and Tours 53316.5.5 Solving ABM 53516.5.6 Integration with Assignment 53616.6 Refining Activity or Tour-Based Models 53716.6.1 Choice of Usual Place of Work and Education 53716.6.2 Car Ownership 53716.6.3 In and Out of Home Activities 53816.6.4 Person Day-Patterns Linked Across Household Members 53816.6.5 Activities Allocated Explicitly Among Members of the Household 53816.6.6 Number of Zones Used 53816.6.7 Time Periods and Time Constraints 53816.6.8 Network Equilibrium 53916.7 Challenges of Activity-Based Models 53916.8 Extending Random Utility Approaches 54017 Model Design 54117.1 Introduction 54117.2 Accuracy and Precision 54217.3 Model Specification 54317.3.1 Model Objectives 54317.3.2 Identify Possible Interventions 54417.3.3 Identify Relevant Behavioural Responses 54417.3.4 Technical Specification and Data Requirements 54517.3.5 Quality Assurance 54617.4 Model Calibration and Validation 54717.5 Model Review 54817.6 Plan Making 54817.7 Dealing with Uncertainty 55018 Key Parameters, Planning Variables, and Value Functions 55318.1 Forecasting Planning Variables 55318.1.1 Introduction 55318.1.2 Use of Official Forecasts 55418.1.3 Forecasting Population and Employment 55518.1.3.1 Trend Extrapolation 55518.1.3.2 Cohort Survival 55518.1.3.3 Transitional Probabilities 55618.1.3.4 Economic Base 55618.1.3.5 Input–Output Analysis 55618.1.4 The Spatial Location of Population and Employment 55718.2 Land-Use Transport Interaction Modelling 55718.2.1 The Lowry Model 55918.2.2 The Bid-Choice Model 56018.2.2.1 Elasticities in Bid-Auction Location Choice Models 56018.2.2.2 Consumer Surplus 56118.2.3 Systems Dynamics Approach 56118.2.4 Urban Simulation 56318.3 Car-Ownership Forecasting 56418.3.1 Background 56418.3.2 Time-Series Extrapolations 56518.3.3 Econometric Methods 56818.3.3.1 The Method of Quarmby and Bates (1970) 56818.3.3.2 The Regional Highway Transport Model (RHTM) Method 57018.3.3.3 Models of Car Ownership and Use 57018.3.3.4 Models of Motorcycle Ownership 57118.3.4 International Comparisons 57118.4 The Value of Travel Time 57318.4.1 Introduction 57318.4.2 Subjective and Social Values of Time 57418.4.3 Some Practical Results 57518.4.4 Methods of Analysis 57618.4.4.1 Estimation of Subjective Values of Time 57618.4.4.2 Confidence Intervals for the Value of Time 57718.4.4.3 A Deeper Look at Computing Measures of Uncertainty for WTP 58018.4.4.4 Special Problems Brought in by the Use of More Flexible Models 58218.4.4.5 The Transfer Price Approach 58918.4.4.6 The Stated Preference Approach 59018.5 Valuing External Effects of Transport 59018.5.1 Introduction 59018.5.2 Methods of Analysis 59118.5.2.2 Contingent Valuation 593Exercises 59619 Pricing and Revenue 59919.1 Pricing and Welfare 59919.2 Correcting Prices for Externalities 60019.3 The Perception of Travel Costs 60119.4 Pricing Tools 60119.4.1 Car Ownership Taxes 60219.4.2 Fuel Taxes 60219.4.3 Parking Charges 60219.4.4 Tolled Facilities 60319.4.5 Pay-As-You-Drive Insurance 60419.4.6 Congestion and Road User Charging 60419.5 The Experience of Private Sector Projects 60519.5.1 Involvement of Private Sector in Transport Projects 60519.5.2 Uncertainty and Risk 60719.5.3 Risk Management and Mitigation 60919.6 Demand Modelling 61019.6.1 Willingness to Pay 61019.6.2 Simple Projects 61019.6.3 Complex Projects 61119.6.4 Road User Charge Projects 61319.6.5 Scheme Design 61319.6.6 Ramp-Up, Leakage, and Discounts 61519.7 Risk Analysis 61619.7.1 Sensitivity and Sources of Risk 61719.7.2 Stochastic Risk Analysis 61819.7.3 Scenarios 61919.8 Conclusions 620Exercises 62020 Modelling the Less Common 62320.1 Introduction 62320.2 New Scheduled Services 62420.3 Walking 62520.4 Cycling 62520.5 Motorcycling 62720.6 Parking 62820.7 Demand-Responsive Transport 63020.7.1 Introduction 63020.7.2 Micro-Mobility Sharing 63320.7.3 Car, Motorcycle, and Van Sharing 63420.7.4 Connected and Automated Vehicles 63520.7.5 Mobility as a Service 63620.8 Modelling Demand-Responsive Mobility 63620.8.1 The Challenge 63620.8.2 Modelling Approaches 63820.8.3 Model Outputs 64020.9 Deliveries and Collections 64020.10 Digital and Distant Presence 64120.11 Soft Measures, Smarter Choices 642References 645Index 689