Why temperature increase during simulated annealing?

I’m using the SA solver in Galapagos for an optimization problem. I found the temperature increased after many generations. However, the temperature should decrease theoretically as the evolution process. Did you come across the same issue? And who knows what’s meaning of the green and blue areas in the plot figure. Thanks a lot.

The temperature does decrease, but after a while the current process terminates and a new one starts, with high temperature again. If the annealing solver didn’t restart it would get stuck in local minima all the time.