A reinsurer rates a satellite with five redundant subsystems and needs the probability the whole craft survives a year. Same math, smaller stakes: a Markov model for policy lapse asks what fraction of a cohort stays in force after ten renewals. Reliability and Markov chains share one engine, the multiplication of independent probabilities arranged into a structure.
Series systems. Components are in series when every one must work for the system to work. With independent components and individual reliabilities ,
Ten components each at 0.99 give . Adding cheap parts to a series chain is expensive in reliability.
KEY: In series, multiply reliabilities. In parallel, multiply unreliabilities. The two operations are duals.
Parallel systems. Components are in parallel when the system works if at least one component works. With independent unreliabilities ,
Three components each at 0.80 give . Redundancy buys reliability cheaply.
Common mistakes
- Multiplying reliabilities in parallel. . The correct rule multiplies unreliabilities: . Two components at 0.80 give 0.96 in parallel, not 0.64.
- Enumerating bridge paths without inclusion-exclusion. Adding the four path probabilities double-counts shared components. Either condition on or apply inclusion-exclusion rigorously.
- Transposing the transition matrix. Some texts use column-stochastic conventions. On MAS-I, rows sum to 1 and distributions multiply from the left.
Bottom line
- Series: . Multiplies survival probabilities, so the weakest link dominates and reliability falls fast as components are added.
- Parallel: . Redundancy multiplies failure probabilities, not survival probabilities.
- Bridge: 5-component diamond with a crossover. Use conditioning on the bridge element for a clean two-line calculation.
- Markov transition matrix : nonnegative entries, rows sum to 1, entry is the one-step probability of moving from state to state ; -step probabilities live in .
Exam shortcut
When a series chain has many components at high reliability, approximate for a quick check; the approximation is tight when individual unreliabilities are under 0.05. For two-state Markov chains with transition probabilities (from 1 to 2) and (from 2 to 1), the stationary distribution is ; memorize this since it appears every sitting.
The full lesson (about 1,888 words, 13 min read) adds 2 worked examples, all 6 common mistakes, a self-check, free in the app.
Learning objectives
- A8
Browse all free MAS-I lessons or jump into free MAS-I practice questions.