Why Do We Usually Take Multiple Reads (Samples)?

The biases and couplings of the qubits define an energy landscape, and the D-Wave™ quantum computer uses quantum annealing to find the minimum energy of that landscape. By the end of the annealing cycle, each qubit is in a classical state that represents the minimum energy state of the problem, or one very close to it. The qubit states are then read-out.

One anneal-read cycle is called a sample.     

Because the annealing time is finite, there is a chance that the system ends up in an excited state. We then have to take multiple samples to find the lowest-energy state.

Taking multiple samples for a specific problem is a way to mitigate the influence of outside energy sources on the quantum system, as no real-world computation can run in perfect isolation.  Certain factors may cause the system to jump from the ground state into a higher energy state.  One is thermal fluctuations that exist in any physical system. 

In reality, for some problems, the probability of staying in the lowest-energy state can sometimes be small; however, the low-energy states that are returned are still very useful.  By taking many samples, we can build a distribution of these possible solutions.

Read more about this in the Introduction to Quantum Annealing section of the Getting Started guide.