================================================================================
## Quick Install

- mkdir build
- cd build
- cmake ..
- make

## Preparation

- Select a CNF file from `data/cnf` as the Boolean formula.
- Select a UAI file from `data/uai` as the probabilistic graphical model. 
- Compile the UAI file with ACE. For example, to run on  `smoker_10.uai` file, we need to use ACE (http://reasoning.cs.ucla.edu/ace) to compile it into `smoker_10.uai.ac` and `smoker_10.uai.lmap` files. 

- Specify the marginalized variable using MAR file. Then we need to specify the variables of the UAI file to be mariginalzed (other variables are shared with the CNF file). MAR files used in our experiments can be found in `data/benchmark_inst/inst_x_x_<cnf_name>_<uai_name>`.

## Example run:

Running command:

```
$koco_smc <cnf_filename> -no-luby -rinc=1.5 -phase-saving=0 -rnd-freq=0.02 -ac <.ac file from ACE> -lmap <.lmap file from ACE> -mar <.mar files from our benchmark SMC instances> -threshold <threshold value> -use-map
```

Example:

```
$koco_smc ./data/example/kcolor_3_10x10.cnf -no-luby -rinc=1.5 -phase-saving=0 -rnd-freq=0.02 -ac ./data/example/smokers_10.uai.ac -lmap ./data/example/smokers_10.uai.lmap -mar ./data/example/mar.mar -threshold 1e137 -use-map
```