MAST30001 Chap.1 Probability Foundations for Stochastic Models
Probability Foundations for Stochastic Models
A stochastic model begins with a system, a state description and an explicit account of randomness. This chapter reviews the probability tools that support the later process models: conditional probability, expectation, distribution functions, exponential waiting times and inverse-transform simulation.
The central habit is to define the random object and its units before calculating, then choose a representation that matches the event being asked about. That discipline prevents category errors, such as treating a density as a probability or confusing an event count with an arrival time.
It also makes later assumptions visible: independence, memorylessness and the Markov property are modelling claims, not algebraic conveniences.
What this chapter covers
- 01
Why physical, mathematical and computational models simplify real systems
- 02
Choosing a state that retains the information needed for prediction
- 03
Events, conditioning and the law of total probability
- 04
Expectation, variance, independence and random sums
- 05
Exponential waiting times and memorylessness
- 06
Inverse-transform simulation from uniform random values
Worked example · free
Simulate a three-point distribution
- 1Cumulative cut-points are 0.20, 0.70 and 1.00.
- 1Map the intervals to outcomes 0, 1 and 2 in order.
- 1The four draws produce 0, 1, 2 and 1.
- 1Their sample mean is 1, which need not equal the model mean 1.10 in a short run.
Key terms
- State Variable
- A quantity or vector that records the information the model uses to determine the future law.
- Conditional Probability
- Probability calculated after specified information is treated as known, changing the relevant sample space.
- Expected Value
- A probability-weighted average that describes the long-run centre of a random quantity.
- Memoryless Distribution
- A waiting-time distribution whose remaining-time law does not depend on the elapsed wait.
- Inverse Transform
- A simulation method that maps a uniform random draw through cumulative probability cut-points.
- Support
- The set of values a random variable can take with positive probability or density under the model.
- Law of Total Expectation
- A rule that computes an expectation by averaging conditional expectations across a complete set of cases.
Probability Foundations for Stochastic Models FAQ
Why must the state be defined before calculating?
The state determines whether the future can be predicted from current information alone and fixes the units and sample paths. Omitting relevant history can destroy the Markov property, while including unnecessary history makes the model harder without improving its predictive law.
When should I condition on another variable?
Condition when the chosen variable splits the experiment into complete, simpler cases. Useful choices include the first transition, the number of arrivals or the current state. Write the conditional law in each case, verify the case probabilities sum to one, then average back.
Why can a simulation disagree with the theoretical mean?
A short simulation is one random sample, so its frequencies and mean fluctuate. Correctness means the generating intervals have the right probability widths and repeated independent runs converge toward the model behaviour, not that every small run reproduces the expectation exactly.
How can I tell whether two random quantities are independent?
Do not infer independence merely because two variables have similar distributions or zero covariance. Use the construction of the experiment, a stated assumption, or a factorisation of their joint law. If they share an underlying arrival, state or time interval, identify that common source before simplifying products or variances.
What is a useful check for an inverse-transform table?
Order outcomes according to their cumulative masses, begin the first interval at zero and end the final interval at one. Every interval width must equal the corresponding probability. Boundary conventions can be chosen consistently because a continuous uniform draw hits any single cut-point with probability zero.
When does linearity of expectation require independence?
It never does. The expectation of a finite sum equals the sum of expectations whenever the expectations exist. Independence becomes relevant for factorising expectations of products and for eliminating covariance terms in variances, so state it only where that stronger operation is actually used.
Exam move
Build a compact probability checklist before starting the process chapters. For each problem, name the experiment, define the random variable, state its support and units, and decide whether conditioning creates smaller copies of the problem. Rehearse converting a cumulative distribution into simulation intervals and checking that the widths sum to one.
Keep independence and linearity separate: expectations can add without independence, while variance calculations may need covariance. Finally, test any waiting-time use of memorylessness by asking whether the source distribution is exponential rather than merely positive. For revision, solve the same short scenario three ways: as an event statement, as a conditional calculation and as a simulation rule.
Compare the inputs each representation needs. When checking an answer, confirm probabilities lie in the unit interval, a cumulative distribution is nondecreasing, simulated interval widths match their masses and every expectation carries the units of the random variable rather than probability units.
Working through Probability Foundations for Stochastic Models in MAST30001? Sia is AskSia’s AI Statistics tutor — ask any MAST30001 Probability Foundations for Stochastic Models question and get a clear, step-by-step explanation grounded in how MAST30001 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
