Home

Multiplicity Automata Learning Library (MALL)

We apply a probably approximately correct learning algorithm for multiplicity automata to generate quantitative models of system behaviors with a statistical guarantee. Using the generated model, we give two analysis algorithms to estimate the minimum and average values of system behaviors. We show how to apply the learning algorithm even when the alphabet is not fixed. The experimental result is encouraging; the estimation made by our approach is almost as precise as the exact reference answer obtained by a brute-force enumeration.

For more detail information, please see our thesis "Quantitative Analysis using Multiplicity Automata Learning."

1. Download
2. Prerequisite & Installation
3. Tutorial
4. Experiment Results

## Download

MALL consists of the following 3 folders:

Folder	Description
Src/main	The core library code in Matlab
Src/experiment	The experiment results presented in the thesis.
Src/enumeration	The code to generate the enumeration results.

You can download the latest version of MALL from git:

git clone https://github.com/fmlab-iis/ma-learning.git

## Prerequisite & Installation

Prerequisites

1. Installed Matlab 2016a with Symbolic Math Toolbox.
2. Perl 5.24.1

Installation

The codes on git are ready to use. You can simply add the path of the directory in the Matlab command window.

## Tutorial

SRC means the path where you clone MALL.

Main

The core function of this learning algorithm is the "order_comp" The function is comprised of the following formation:

[timeout, time_each, state, counter_mem, mem,eq_mem] = order_comp( order, distri, probability )

Experiment

Open the Matlab

For the result of Table1, type the following command in Matlab command line:

cd('SRC/experiment/Table1') 
experiment

For the result of Table2, type the following command in Matlab command line:

cd('SRC/experiment/Table2') 
experiment

For the result of Table3, type the following command in Matlab command line:

cd('SRC/experiment/Table3') 
experiment

For the result of Table4, type the following command in Matlab command line:

cd('SRC/experiment/Table4') 
experiment

For the result of Table5, type the following command in Matlab command line:

cd('SRC/experiment/Table5') 
experiment

For the result of Table6, type the following command in Matlab command line:

cd('SRC/experiment/Table6') 
experiment

For the result of Table7, type the following command in Matlab command line:

cd('SRC/experiment/Table7') 
experiment

Enumeration

For the result of enumeration of Table 5, type the following command in Matlab command line:

cd('SRC/enumeration/Table5') 
os_oracle

For the result of enumeration of Table 6, type the following command in Matlab command line:

cd('SRC/enumeration/Table6') 
mc_oracle

For the result of enumeration of Table 7, type the following command in Matlab command line:

cd('SRC/enumeration/Table7') 
web_oracle

## Experiment Results

Here are some experiment results and detail information of the tables.

Table1

The following table is the experiment result of prediction of the Calculator experiment.

Table2

The following table is the experiment result of the Calculator experiment.

Table3

The following table is the experiment result of different distribution in the Calculator experiment.

Table4

The following table is the experiment result of Calculator experiment with different alphabet symbol size.

Table5

The following table is the Operating System Scheduling experiment.

Table6

The following table is the Ｍissionary & Cannibals experiment.

Table7

The following table is the Amount of Data Transmission in a Website experiment.