Amplitude amplification increases the probability of measuring one or more desired quantum outcomes. It prepares a starting state with A, then applies a Grover iterate G repeatedly: GjA. Each iterate applies an oracle that tags good outcomes by multiplying their amplitudes by −1, followed by an inversion operator that raises good magnitudes while lowering bad ones. section 6.1 section 6.3
Think of the oracle as a predicate over an unordered list encoded as quantum outcomes. Rather than returning the matching item directly, it marks matching outcomes. Inversion redistributes amplitudes so measurement is more likely to return a marked result. The technique is useful only when the oracle can be implemented efficiently; changing amplitudes alone provides no search advantage. section 6.1 section 6.2
A practical implementation follows this pattern:
- Prepare an n-qubit starting state with A, often a uniform distribution over outcomes. section 6.1 section 6.2.4
- Define an oracle O that tags every good outcome. section 6.3
- Build G from O and an inversion operator M; for a state prepared by A, M can be expressed as AM0A−1. section 6.1.5
- Choose an iteration count, apply G repeatedly, then measure. Too many iterations can reduce the desired amplitude again. section 6.1.4
Three-qubit marked outcomes
For eight uniformly prepared outcomes, mark outcomes 1, 3, and 7 with a phase oracle. The iteration estimate selects one Grover iterate, after which reports can inspect the amplitudes of those marked outcomes. A related single-target simulation marks outcome 3 and shows its amplitude increasing after one and then two iterations. section 6.2.4 section 6.1.4
Figure 6.17 The combined measurement probabilities of good and bad outcomes before and after j applications of the Grover iterate
Compare the combined probabilities of good and bad outcomes before and after repeated Grover iterations. section 6.1.4
The same pattern supports optimization: mark selections meeting value and weight criteria, amplify them, and measure candidate solutions. In a knapsack example, amplification raises the probability of selections satisfying a maximum weight and minimum value, though several qualifying selections can still be measured. section 2.3.2 section 2.5
