Drag a stop or the depot to move it; the solver re-optimises live. Use the tool menu to add or remove customers.
Best cost-
Distance-
Total lateness-
Vehicles used-
vs nearest-neighbour-
Temperature-
Cost of the current (grey) and best (green) plan against solver iterations; the red dashed line (when in range) is the nearest-neighbour baseline.
Capacitated vehicle routing with soft time windows (Solomon 1987). Cost = distance + lambda x lateness + 40 x overload + fixed cost per vehicle used. A plan is improved by simulated annealing (Kirkpatrick et al. 1983) using relocate, swap and 2-opt moves; worse plans are accepted with probability exp(-delta/T) while T cools. Red stops are served late; trucks drive the best plan on a 06:00 clock.