Algorithm for allocating tables based on preferences, using simulated annealing

• An installation of Go: https://go.dev/dl/

• go install github.com/mhbardsley/table-allocations@latest
• Create a JSON file to hold people, their preferences and table capacities. See sample.json as an example. (Note: the program will, by default, look for a input.json file)

• table-allocations [flags]
• Note: there may be a small delay between running the program and getting the result (see below for inspecting flags to change speed/optimisation trade-off)

All flags are optional and most do not need touching. If you have not named your JSON file input.json, you need to supply an -f flag, e.g. table-allocations -f sample.json will carry out the algorithm on the sample data.

Another useful flag is -m, which specifies what is being optimised. There are three options: sum, which will optimise the total number of preferences satisfied; count, which will optimise the number of people with at least 1 satisfied preference; hybrid (default), which aims to compromise between these. To choose sum, for example, use table-allocations -m sum.

For all other flags (which don't really need tweaking), you can run with the -h flag, i.e. table-allocations -h.