Here are the implementations of algorithms solving the Longest Path problem. Gradient descent, Metropolis and Simulated Annealing are implemented.
Repository content is for Algorithms and Data Structures II course homework in NRU HSE
Algorithms results are represented on this graph. On horizontal line there are different 100-vertices graphs and on the vertical one there are average results of each algorithm working on the current graph. As we can see Gradient descent algorithm gives us the worst result (the blue line is usually below two others). Metropolis and Simulation annealing are pretty much the same. This tendency is evident because Gradient descent can never go away from local minima (maxima in this case). And for Metropolis algorithm there is always a probability to escape local extrema.