Question 1 of 7: Trip Frequency, Travel Demand Management and Highway Capacity

Question 1: Trip Frequency, Travel Demand Management and Highway Capacity (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Part (a) — Socio-economic characteristics and trip frequency

Trip frequency — the number of trips a household or a person makes in a day — is the output of the trip generation step of the four-step model, and it is explained almost entirely by household socio-economic variables rather than by anything about the transport network. The behavioural logic is that travel is a derived demand: a household travels because its members must reach activities, so the number of trips follows the number of people who have activities to reach and the household's ability to reach them.

The variables that carry the explanatory power, in roughly descending order of importance in Canadian household travel surveys, are these. Household size is the single strongest predictor, because each additional member brings their own activity programme; a five-person household makes roughly twice the daily trips of a two-person household, though not five-halves as many, because trips are shared (one parent chauffeurs two children on one chain). Number of employed members matters more per person than household size does, because each worker contributes a compulsory, peak-period, round-trip commute plus the work-based sub-tours that hang off it. This paper's own Question 3 quantifies exactly that contrast: the fitted rate Trip rate = 0.2 + 0.5 NRES + 1.1 NWOR gives each additional worker (1.1 trips/day) 2.2 times the weight of each additional resident (0.5 trips/day). Household income raises trip frequency because discretionary travel (shopping, recreation, personal business) is income-elastic, and because income buys the vehicle that makes those trips cheap. Automobile ownership is the strongest single proxy in most calibrated models — a zero-car household in a Canadian suburb makes on the order of half the daily trips of a two-car household — although it is partly endogenous, since households that intend to travel a great deal buy cars. Life-cycle stage (age of the head, presence and age of children, retirement) reorganises both the number and the purpose mix: a household with school-age children generates the serve-passenger trips that dominate the school peak, while a retired couple makes fewer but more midday, more discretionary trips. Driver's-licence holding, gender roles and dwelling type round out the list; dwelling type is partly a land-use variable in disguise, since apartment households in a dense core substitute walk trips for car trips.

Two cautions belong in a professional answer. First, these variables are strongly collinear — income, auto ownership and household size move together — so a regression that includes all of them will show unstable coefficients and must be specified carefully or replaced by a cross-classification model. Second, the socio-economic variables explain the number of trips well but say almost nothing about where or how those trips are made; destination and mode are governed by accessibility and level of service, which is why the four-step model separates generation from distribution, mode split and assignment.

Part (b) — A travel demand management strategy and its social impacts

Travel demand management (TDM) is the family of measures that changes travel behaviour rather than adding capacity: it works on the demand side of the market by altering the generalised cost of the alternatives. A concrete, quantified example is available in this very paper. Question 7 asks for the mode shares in a four-mode corridor before and after a city builds a bike path that cuts cycling time from 60 min to 45 min. Solving the multinomial logit model there gives a cycling share that rises from 24.8 per cent to 37.5 per cent, with automobile falling from 36.1 to 30.0 per cent, rail from 27.4 to 22.8 and bus from 11.8 to 9.8 per cent — a 6.1-point reduction in car mode share bought with infrastructure that costs a small fraction of a lane of urban arterial.

Other measures of the same family, and the lever each pulls, are worth naming: congestion or cordon pricing and parking pricing (raise the out-of-pocket cost term TC for the car); employer transit-pass subsidies and the federal/provincial tax treatment of transit benefits (lower TC for transit); flexible and staggered work hours and telework (remove the trip or move it out of the peak, changing the temporal distribution rather than the daily total); high-occupancy-vehicle and bus-priority lanes (lower riding time RT for the shared modes and raise it for the single-occupant car); and ride-matching and car-share programmes (raise average vehicle occupancy so that person-trips are served with fewer vehicle-trips).

The social consequences are mixed and must be stated as such. On the benefit side: fewer vehicle-kilometres travelled means lower greenhouse-gas and criteria-pollutant emissions, fewer collisions, and deferred capital spending on road widening; a modal shift toward active transport delivers a public-health dividend through physical activity; and relieving the peak improves the reliability that freight and emergency services depend on. On the cost side: pricing measures are regressive unless the revenue is recycled, because a flat toll is a far larger share of a low-income household's budget; a mode shift away from the car can reduce the accessibility of suburban and rural households who have no realistic alternative; retailers on a priced or reallocated corridor commonly resist the loss of kerbside parking; and induced demand can claw back part of the relief, since travel time freed on the road network is partly refilled by trips that were previously suppressed. A defensible TDM programme therefore combines a push measure with a pull measure, phases them so that the alternative exists before the penalty bites, and recycles revenue into the alternatives — which is the design used in the Metro Vancouver and Greater Toronto regional transportation plans.

Part (c) — The effect of three factors on highway capacity

Given. A basic freeway segment with an ideal capacity of about 2,300 pc/h/ln (HCM 6th ed., 100 km/h free-flow speed); illustrative heavy-vehicle percentage P_T = 15 per cent, passenger-car equivalent E_T = 2.0 on level terrain and E_T = 4.5 on a specific steep grade. Find. The direction — increase or decrease — in which each of the three listed factors moves capacity, with the mechanism named and the magnitude illustrated where it can be computed.

Higher percentage of trucks and buses — capacity DECREASES. Capacity is measured in vehicles per hour but is governed by the space each vehicle occupies in the traffic stream. A heavy vehicle is longer, accelerates and decelerates more slowly, and is driven with a larger following headway, so it consumes the road space of several passenger cars; it also cannot hold speed on a grade, which forces passing manoeuvres and increases speed variance in the adjacent lanes. The HCM handles this with the heavy-vehicle adjustment factor

so a stream with 15 per cent heavy vehicles on level terrain carries about 13 per cent fewer vehicles per hour than the same facility with cars only — roughly 2,300 down to 2,000 veh/h/ln. The effect is proportional to the heavy-vehicle share and grows sharply with grade.

More lanes — capacity INCREASES. Directional capacity is the per-lane capacity multiplied by the number of lanes, so the first-order effect is simply proportional: a four-lane directional cross-section carries about twice the flow of a two-lane one, some 9,200 pc/h against 4,600 pc/h at 2,300 pc/h/ln. There is a smaller second-order gain as well: the HCM free-flow-speed model raises the estimated free-flow speed as the number of lanes increases, because drivers on a wider cross-section have more opportunity to pass and are less constrained by the slowest vehicle, so per-lane capacity itself edges upward. The gain is not unlimited — weaving, lane-changing turbulence and the capacity of the downstream bottleneck eventually govern, and adding lanes on a congested corridor induces additional demand that erodes part of the benefit.

Steeper downgrade — capacity DECREASES. This is the sub-part candidates most often get wrong, because the intuition that "gravity helps" applies only to passenger cars, whose capacity is essentially unaffected by a downgrade. The governing mechanism is again the heavy vehicle: a loaded truck descending a long, steep grade is speed-limited by its ability to dissipate braking energy without overheating, so it is driven well below the passenger-car speed and generates the same platooning and speed variance that an upgrade does. The HCM therefore publishes passenger-car equivalents for specific downgrades that exceed unity; at E_T = 4.5 for a long steep descent,

a capacity reduction of about 34 per cent at the same 15 per cent heavy-vehicle share. On a downgrade shallower than roughly 3 per cent, or on a facility carrying negligible truck traffic, the effect is small enough to ignore — so the honest statement is that a steeper downgrade reduces capacity, and that the reduction is mediated entirely by heavy vehicles.