Releases: nick-falco/eat
Releases · nick-falco/eat
v0.0.3 - Parallel BEAM algorithm
An implementation of BEAM algorithm as described in the paper EAT 3. For a given beam, processes will run at the max full beam height H until the next beam level H+1 is full. Once the next beam level H+1 is completely full, all processes running at level H are terminated and reassigned to run at level H+1. The algorithm uses the Gambler Ruin Algorithm (GRA) to randomly generate candidate female terms.
- Adds support for running the BEAM algorithm with searching for valid female terms at a given height performed in parallel
- Adds options for tuning BEAM algorithm parallelism: --beam-width (-bw) and --beam-timeout (-bto)
- Adds valid female term search time to BEAM algorithm verbose output
- Makes printing validity arrays in verbose output optional. You can now tell the program to include the validity array by including the (-iva / --include-validity-array) parameter.
Support target array containing a list of values
- Add support for specifying a target array containing a list of values
- Add support for specifying the number of values in the target array that should be forced to accept any possible output value
Beam Enumeration Algorithm
- Adds Beam Enumeration Algorithm to the list of supported algorithms
Converging Beam Algorithm with an incorrect output
- Adds a version of the Beam algorithm that converges, but produces an incorrect result
- Adds option to pass a random term growth probability for the Gamblers Ruin Algorithm to use
CLI with Deep Drilling Algorithm
- A CLI with support for running a simple implementation of the Deep Drilling Algorithm