Optimization Lab
Annealing Simulator
Explore how temperature controls stochastic search. Accepted uphill moves help the solver escape local minima early, while cooling gradually turns exploration into exploitation.
Landscape
Two basins with one global minimum and one tempting local minimum.
Step 160; temperature 0.008; current energy -2.902; best -2.902.
Learning mode
Escape the false valley
Annealing is hiking in fog while the weather cools: early wandering helps escape valleys, later steps settle.
Raise initial temperature, slow the cooling rate, and scrub the trace slider.
Definitions on demand
Temperature
A search parameter controlling tolerance for uphill energy moves.
Acceptance
Whether a proposed move becomes the next state.
Cooling schedule
The rule that lowers temperature as the run progresses.
Energy Landscape
Delta E
0.022
Temperature
0.008
P(accept)
0.074
Result
Rejected
Best Energy Trace