Questions & explanations
1. Give an example where user equilibrium differs from system optimum.
Consider two routes: Route 1 takes 10 minutes + 1 minute per 100 cars, Route 2 takes 20 minutes + 1 minute per 100 cars. With 1000 cars, user equilibrium splits 500/500, giving each 15 minutes (10+5 and 20+5). System optimum assigns 600 to Route 1 (10+6=16 min) and 400 to Route 2 (20+4=24 min), total time 600*16+400*24=19200 vs user equilibrium 1000*15=15000? Actually user equilibrium total is 15000, system optimum total is 19200? That's higher—let's correct: For system optimum, we minimize total time. With 600 on Route 1, time=16, 400 on Route 2, time=24, total=600*16+400*24=9600+9600=19200. With 500 each, total=500*15+500*25=7500+12500=20000. So system optimum gives lower total time (19200 < 20000). So system optimum assigns more to the faster route to reduce overall congestion.
2. What is the transit network design problem (TNDP)?
The transit network design problem (TNDP) is about choosing bus or train routes and their schedules to best serve passengers while keeping costs low. It is a complex optimization problem because many factors like demand, travel time, and budget must be balanced. Metaheuristics are general solution methods that find good-enough answers when exact solutions are too hard. Genetic algorithms are one type of metaheuristic that mimics natural selection: they start with many random route plans, keep the best ones, and combine them to create better plans over many steps. Other metaheuristics include simulated annealing and particle swarm optimization. These methods help planners design efficient transit networks without testing every possible option.
3. How does a genetic algorithm solve the TNDP?
A genetic algorithm solves the TNDP by treating each possible route plan as an 'individual' in a population. It starts with many random plans, then evaluates each plan using a fitness function that measures how well it serves passengers and uses resources. The best plans are selected to 'breed' by combining parts of two plans to create new plans, and sometimes random changes (mutations) are added. This process repeats for many generations, gradually improving the population. Eventually, the algorithm produces a very good route plan, though not necessarily the perfect one. Genetic algorithms are popular because they can handle many constraints and find good solutions in reasonable time.
4. What is shared mobility?
Shared mobility means services like bike-sharing, scooter-sharing, and ride-hailing that let people use a vehicle for a short time without owning it. These services help people travel from their home or work to a bus stop or train station, which is called the first or last mile. When shared mobility is connected to public transit, it makes the whole trip easier. But cities face challenges like making sure bikes and scooters are parked safely and not blocking sidewalks. They also need rules to keep ride-hailing from causing more traffic. Good integration means you can use one app to plan and pay for both the shared ride and the transit trip.
5. How can a MaaS app use A/B testing to improve user onboarding?
User onboarding is the first experience when someone downloads the app. A MaaS app could test two different onboarding flows: one that asks for preferences immediately and another that lets users explore first. The metric could be 'percentage of users who complete a trip within the first week.' If the preference-based flow leads to more trips, it is better. The app could also test different welcome messages or tutorial lengths. A/B testing helps find the most effective way to get new users to start using the service. This reduces the number of users who abandon the app after first launch. Good onboarding increases long-term retention.
6. Compare rule-based chatbots and NLP-based chatbots for MaaS customer support.
Rule-based chatbots follow fixed rules like 'if user says 'bus', answer with bus schedule.' They are simple and cheap but can only handle limited questions. If a user asks something unexpected, the chatbot fails. NLP-based chatbots use machine learning to understand a wider range of phrases and intents. For example, they can understand 'What time does the 42 bus come?' and 'When is the next 42?' as the same question. NLP chatbots learn from data and improve over time. However, they are more complex to build and require training data. In MaaS, a hybrid approach often works best: NLP for understanding and rules for specific actions.
7. Compare synchronous and asynchronous communication between microservices in MaaS.
Synchronous communication means one service waits for an answer from another before continuing. For example, when booking a ride, the booking service waits for payment confirmation. This is simple but can slow things down if the other service is busy. Asynchronous communication means the first service sends a request and moves on without waiting. Later, it gets the answer. In MaaS, a route planner might send a request to many transport providers at once and then collect their replies as they come. Asynchronous is faster and more resilient because a slow provider doesn't block everything. However, it is harder to design and debug.
8. What is the difference between a metaheuristic and an exact optimization method for TNDP?
An exact optimization method guarantees finding the best possible route plan, but it can take very long for large problems because it checks all possibilities. Metaheuristics, like genetic algorithms or simulated annealing, do not guarantee the absolute best plan but find a very good one much faster. They are used when the problem is too big for exact methods. For example, a city with hundreds of bus stops would take too long to solve exactly, so planners use metaheuristics. The trade-off is speed versus optimality: metaheuristics are practical for real-world planning, while exact methods are used for small or theoretical cases.
9. What is the main goal of universal design in MaaS?
Universal design in MaaS means making the service easy to use for everyone, including people with disabilities, older adults, and those who are not good with technology. The main goal is to remove barriers so that all people can plan and pay for trips using different transport modes like buses, trains, and bike-share. For example, an app should work with screen readers for blind users and have simple buttons for people who find complex menus hard. It also means offering different ways to pay, like cash or card, for people without smartphones. By focusing on universal design, MaaS can be fair and useful for the whole community.
10. What was the first form of public transit in cities?
The first form of public transit in cities was the horse-drawn omnibus, which was a large carriage pulled by horses that could carry many passengers along a fixed route. This started in the 1820s in Paris and other cities. Later, horse-drawn streetcars on rails made the ride smoother and allowed horses to pull heavier loads. These systems were slow and limited by the need for many horses and their care. The invention of electric streetcars in the late 1800s replaced horse-drawn transit, making it faster, cleaner, and able to carry more people. This was a key milestone that led to the growth of modern public transit systems.
11. What is cloud computing and why is it important for MaaS?
Cloud computing means using remote servers on the internet to store data and run software instead of using your own computer. For MaaS, the cloud provides the power to handle many users at once, store trip histories, and run complex calculations like route optimization. It is important because MaaS needs to scale up quickly during rush hour and scale down at night. The cloud also offers services like databases and machine learning that MaaS apps use. Without the cloud, a MaaS company would have to buy and manage its own servers, which is expensive and less flexible. Cloud computing makes MaaS affordable and reliable.
12. What is user profiling in a MaaS app?
User profiling means collecting information about a user to understand their travel habits and preferences. For example, the app might note that a user often takes the bus to work at 8 AM and uses ride-hailing late at night. This profile helps the app offer personalized suggestions, like the fastest route or a discount on a frequently used service. Profiling uses data like trip history, payment method, and favorite destinations. It can also include demographics like age or location. The goal is to make the app more useful and convenient for each user. However, it must be done with user permission and respect privacy.