How can a queue occur if utilization is less than $100\%?$ It's because utilization is measured across some time interval. A queue forms when, at some point during this time interval, the resource is already busy $($or all resources are already busy$)$ and more work arrives. Then the new work must wait in a queue until the resource$($s$)$ finish$($es$)$ the old work. But at other times during the time interval, the resource$($s$)$ are not always busy.

Let's say that utilization is less than $100\%$ across some time interval. What is the utilization of the resource$($s$)$ during that portion of the time interval that a queue exists?