Multiple$-$Processor Scheduling

$[$Lisonbee$]$ A system that has multiple cores needs a way to balance work among all of its cores $($load balancing$),$ but at the same time it should be intelligent in the way it handles jobs on each processor in order to avoid potential cache misses. Given this system, propose how the concept of process affinity can be used to optimize for both load balancing as well as the number of cache hits.