It's worth first explaining what open loop control system is.

This is a control system where you essentially blindly set an value. This may work well, based on experience. You may for instance know that a sausage is warm after 10 minutes in warm water bath, without actually checking the internal temperature of the sausage. For some processes, this is OK, because it's either cheap to overrun a little bit, or cheap to underrun.
An example of this would be my home ventilation; I don't check humidity, temperature or any other factors; it gets turned on at 09:00, and off at 22:00, no matter what. The cost of running it is tiny, and the cost of sensors and programming high.
This is also used in most washing machines; they run based on a time, not any measurement of the process value, which is how dirty the clothes are.
In short; a open loop system does not look at the system to determine where the process is. It blindly sets the set point.
Imagine a sled, on a rail, which accepts commands to go left or right for some distance, e.g. go 10 cm left. If you are at the leftmost position, and send this command, it will lead to a crash. In such a situation, open loop system will not suffice.
A closed loop on the other hand, has a feedback.
The Process Value (PV) is fed back into the controller, modifying the set point. Different types of controllers exists; for analogue values, a PID controller is commonly used.
But let's continue with the previous example of the sled on rails. Now the controller knows where the sled is, so the set point to the controller can be changed from go left by 10 cm to go to five centimeters right of the left end stop, and the controller will know which direction it should drive the sled, and for how far.
Another example is the simple thermostat on an oven. The oven (actuator) heats up the room, and the process value is sensed by the thermostat, which turns off the actuator when the setpoint matches the process value.
Control engineering is a big topic, and this is a very brief introduction.