For running the code, you should must to make sure your python has the tkinter module. After that, you can test the result by running the following command:
python3 demo.py
You can pass specific arguments to define parameters, like:
--algorithm : Defines which algorithm should be used. Can be "self" or "tarry". The default value is self.
--agents : Defines the number of agents that will be in the graph. The default value is 3.
--graph : Defines the file of a graph that can be used to generate the graph. The default value is None.
As an example, you could run:
python3 demo.py --agents 3 --graph test1.csv
python3 demo.py --agents 3 --graph full_tree_test.graphml
python3 demo.py --agents 3 --graph test2.csv
python3 demo.py --agents 3 --graph testimperfect.csv
python3 demo.py --agents 3 --graph testperfect.csv
python3 demo.py --agents 3 --graph test1.csv --algorithm two_interval
python3 demo.py --agents 3 --graph test1.csv --algorithm tarry
python3 demo.py --agents 3 --graph test1.csv --algorithm tarry_interval_priority
python3 demo.py --agents 3 --graph test1.csv --algorithm tarry_interval_tie_breaker
python3 demo.py --agents 3 --graph test1.csv --algorithm tarry_delayed_interval_tie_breaker
python3 demo.py --agents 3 --graph "mazes/10_by_10/maze_10x10__1.csv"
python3 demo.py --agents 3 --graph "mazes/20_by_20/maze_20x20__1.csv"
python3 demo.py --agents 3 --graph "mazes/30_by_30/maze_30x30__1.csv"
python3 demo.py --agents 3 --graph "mazes/40_by_40/maze_40x40__1.csv"
python3 demo.py --agents 3 --graph "graphs/random_unlabeled_tree/100/graph_100__1.graphml" --algorithm tarry_delayed_interval_tie_breaker
python3 demo.py --agents 3 --graph "graphs/random_unlabeled_tree/250/graph_250__1.graphml" --algorithm tarry_delayed_interval_tie_breaker
python3 demo.py --agents 3 --graph "graphs/random_unlabeled_tree/500/graph_500__1.graphml" --algorithm tarry_delayed_interval_tie_breaker
python3 demo.py --agents 3 --graph "graphs/random_unlabeled_tree/1000/graph_1000__1.graphml" --algorithm tarry_delayed_interval_tie_breaker
python3 demo.py --agents 3 --graph graphs/barabasi_albert/100/graph_100__190.graphml --algorithm tarry_interval_priority
After making sure the simulation works, you can automatically generate results for multiple sets of graphs. This can take a long time.
You can pass specific arguments to define parameters, like:
--algorithm : Defines which algorithm should be used. Can be "self", "two_interval" or "tarry". The default value is self.
--base_path : Defines the path to the directory where all graphs as stored.
--graph_size_path : Defines specific size directory to use (e.g., '10_by_10', '20_by_20'). If not provided, all sizes found will be used.
--max_agents : Defines the maximum number of users to be simulated.
--plot: Defines if the plot should be generated.
As an example, you could run:
python3 results.py --base_path mazes --algorithm self --max_agents 40 --plot true
python3 results.py --base_path mazes --algorithm two_interval --max_agents 40 --plot true
python3 results.py --base_path mazes --algorithm tarry --max_agents 40 --plot true
python3 results.py --base_path mazes --max_agents 40
python3 results.py --base_path mazes --algorithm tarry_interval_priority --graph_size_path 10_by_10 --max_agents 40 --plot true
python3 results.py --base_path mazes --algorithm tarry_interval_priority --max_agents 40 --plot true
python3 results.py --base_path mazes --algorithm tarry_interval_tie_breaker --graph_size_path 10_by_10 --max_agents 40 --plot true
python3 results.py --base_path mazes --algorithm tarry_interval_tie_breaker --max_agents 40 --plot true
python3 results.py --base_path mazes --algorithm tarry_delayed_interval_tie_breaker --graph_size_path 10_by_10 --max_agents 40 --plot true
python3 results.py --base_path mazes --algorithm tarry_delayed_interval_tie_breaker --max_agents 40 --plot true
python3 results.py --base_path graphs/random_unlabeled_tree --max_agents 40 --plot true
python3 results.py --base_path graphs/random_unlabeled_tree --graph_size_path 100 --max_agents 40 --plot true
python3 results.py --base_path graphs/random_unlabeled_tree --graph_size_path 250 --max_agents 40 --plot true
python3 results.py --base_path graphs/random_unlabeled_tree --graph_size_path 500 --max_agents 40 --plot true
python3 results.py --base_path graphs/random_unlabeled_tree --graph_size_path 1000 --max_agents 40 --plot true
python3 results.py --base_path graphs/random_unlabeled_tree --graph_size_path 1500 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm self --graph_size_path 100 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm two --graph_size_path 100 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry --graph_size_path 100 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_interval_priority --graph_size_path 100 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_interval_tie_breaker --graph_size_path 100 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_delayed_interval_tie_breaker --graph_size_path 100 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --graph_size_path 250 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm self --graph_size_path 250 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm two_interval --graph_size_path 250 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry --graph_size_path 250 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_interval_priority --graph_size_path 250 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_interval_tie_breaker --graph_size_path 250 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_delayed_interval_tie_breaker --graph_size_path 250 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --graph_size_path 500 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm self --graph_size_path 500 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm two_interval --graph_size_path 500 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry --graph_size_path 500 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_interval_priority --graph_size_path 500 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_interval_tie_breaker --graph_size_path 500 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_delayed_interval_tie_breaker --graph_size_path 500 --max_agents 40 --plot true
python3 results.py --base_path graphs/barabasi_albert --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/barabasi_albert --algorithm self --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/barabasi_albert --algorithm two_interval --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_interval_priority --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_interval_tie_breaker --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/barabasi_albert --algorithm tarry_delayed_interval_tie_breaker --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --graph_size_path 100 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry --graph_size_path 100 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_interval_priority --graph_size_path 100 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_interval_tie_breaker --graph_size_path 100 --max_agents 40 --plot
python3 results.py --base_path graphs/small_world --graph_size_path 250 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --algorithm self --graph_size_path 250 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --algorithm two_interval --graph_size_path 250 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry --graph_size_path 250 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_interval_priority --graph_size_path 250 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_interval_tie_breaker --graph_size_path 250 --max_agents 40 --plot
python3 results.py --base_path graphs/small_world --algorithm tarry_delayed_interval_tie_breaker --graph_size_path 250 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --graph_size_path 500 --max_agents 40 --plot false
python3 results.py --base_path graphs/small_world --algorithm self --graph_size_path 500 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm two_interval --graph_size_path 500 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry --graph_size_path 500 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_interval_priority --graph_size_path 500 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_interval_tie_breaker --graph_size_path 500 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_delayed_interval_tie_breaker --graph_size_path 500 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm self --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm two_interval --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_interval_priority --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_interval_tie_breaker --graph_size_path 1000 --max_agents 30 --plot false
python3 results.py --base_path graphs/small_world --algorithm tarry_delayed_interval_tie_breaker --graph_size_path 1000 --max_agents 30 --plot false
After generating results, you can generate specific plots to compare different algorithms.
You can pass specific arguments to define which and how they should be compared, like:
--algorithms : Defines which algorithm should be compared.
--graph_size_path : Defines specific size directory to use (e.g., '10_by_10', '20_by_20'). If not provided, all sizes found will be used.
--metrics : DefineS which metric should be used. Defaults to ALL.
--split_legend : Defines if the legend should be generated alongside the plot or separate.
--title : Defines if the title for the plot.
As an example, you could run:
python3 compare.py --algorithms "self, two_interval, tarry" --base_path mazes --graph_size 40_by_40 --metrics "avg_pioneer_steps" --split_legend False --title "Algorithms Pioneer Steps - Comparison"
python3 compare.py --algorithms "self, two_interval, tarry" --base_path mazes --graph_size 40_by_40 --metrics "avg_fraction_pioneer" --split_legend False --title "Algorithms Pioneer Fraction - Comparison"
python3 compare.py --algorithms "tarry,tarry_interval_priority,tarry_interval_tie_breaker" --base_path mazes --graph_size 40_by_40 --metrics "avg_pioneer_steps" --split_legend True --title "Tarry Variants - Comparison"
python3 compare.py --algorithms "tarry,tarry_interval_priority,tarry_interval_tie_breaker, tarry_delayed_interval_tie_breaker" --base_path mazes --graph_size 40_by_40 --metrics "avg_pioneer_steps" --split_legend True --title "Tarry Variants - Comparison"
python3 compare.py --algorithms "self, two_interval, tarry" --base_path graphs/random_unlabeled_tree --graph_size 100 --metrics "avg_pioneer_steps" --split_legend False --title "Algorithms Pioneer Steps Random Unlabeled - Comparison"
python3 compare.py --algorithms "tarry,tarry_interval_priority,tarry_interval_tie_breaker, tarry_delayed_interval_tie_breaker" --base_path graphs/random_unlabeled_tree --graph_size 100 --metrics "avg_pioneer_steps" --split_legend True --title "Tarry Variants Random Unlabeled - Comparison"
python3 compare.py --algorithms "self, two_interval, tarry" --base_path graphs/barabasi_albert --graph_size 100 --metrics "avg_pioneer_steps" --split_legend True --title "Algorithms Pioneer Steps - Comparison - 100 Barabási-Albert Graph"
python3 compare.py --algorithms "self, two_interval, tarry" --base_path graphs/barabasi_albert --graph_size 100 --metrics "avg_fraction_pioneer" --split_legend True --title "Algorithms Pioneer Fraction - Comparison - 100 Barabási-Albert Graph"
python3 compare.py --algorithms "tarry,tarry_interval_priority,tarry_interval_tie_breaker, tarry_delayed_interval_tie_breaker" --base_path graphs/barabasi_albert --graph_size 1000 --metrics "avg_pioneer_steps" --split_legend True --title "Tarry Variants Pioneer Steps- Comparison - 1000 Barabási-Albert Graph"
python3 compare.py --algorithms "tarry,tarry_interval_priority,tarry_interval_tie_breaker, tarry_delayed_interval_tie_breaker" --base_path graphs/barabasi_albert --graph_size 1000 --metrics "avg_fraction_pioneer" --split_legend True --title "Tarry Variants Pioneer Fraction- Comparison - 1000 Barabási-Albert Graph"
python3 compare.py --algorithms "self, two_interval, tarry" --base_path graphs/small_world --graph_size 100 --metrics "avg_pioneer_steps" --split_legend True --title "Algorithms Pioneer Steps - Comparison - 100 Small World Graph"
python3 compare.py --algorithms "self, two_interval, tarry" --base_path graphs/small_world --graph_size 100 --metrics "avg_fraction_pioneer" --split_legend True --title "Algorithms Pioneer Fraction - Comparison - 100 Small World Graph"
python3 compare.py --algorithms "tarry,tarry_interval_priority,tarry_interval_tie_breaker, tarry_delayed_interval_tie_breaker" --base_path graphs/small_world --graph_size 100 --metrics "avg_pioneer_steps" --split_legend True --title "Tarry Variants Pioneer Steps- Comparison - 100 Small World Graph"
python3 compare.py --algorithms "tarry,tarry_interval_priority,tarry_interval_tie_breaker, tarry_delayed_interval_tie_breaker" --base_path graphs/small_world --graph_size 100 --metrics "avg_fraction_pioneer" --split_legend True --title "Tarry Variants Pioneer Fraction- Comparison - 100 Small World Graph"
python3 compare.py --algorithms "tarry,tarry_interval_priority,tarry_interval_tie_breaker, tarry_delayed_interval_tie_breaker" --base_path graphs/small_world --graph_size 100 --metrics "avg_pioneer_steps" --split_legend True --title "Tarry Variants Pioneer Steps- Comparison - 100 Small World Graph" --relative True
For a deeper analysis, we can calculate statistics related to the graphs used.
--graph-path : Defines which file path to saved graph(file) or saved graphs(directory) should be analysed.
As an example, you could run:
python3 graph_statistics.py --graph-path "test1.csv"
python3 graph_statistics.py --graph-path "mazes/10_by_10"
python3 graph_statistics.py --graph-path "mazes/10_by_10/maze_10x10__1.csv"
python3 graph_statistics.py --graph-path "mazes/20_by_20"
python3 graph_statistics.py --graph-path "mazes/20_by_20/maze_20x20__1.csv"
python3 graph_statistics.py --graph-path "mazes/30_by_30"
python3 graph_statistics.py --graph-path "mazes/30_by_30/maze_30x30__1.csv"
python3 graph_statistics.py --graph-path "mazes/40_by_40"
python3 graph_statistics.py --graph-path "mazes/40_by_40/maze_40x40__1.csv"
python3 graph_statistics.py --graph-path "graphs/random_unlabeled_tree/100"
python3 graph_statistics.py --graph-path "graphs/barabasi_albert/100"
python3 graph_statistics.py --graph-path "graphs/small_world/100"
You can also create your own dataset of graphs.
--graph_type : Defines which generator should be used for creating the graphs.
--graph_size : Defines the size of the graph(number of nodes).
--n_graph : Defines how many graphs should be generated.
As an example, you could run:
python3 generate_graphs.py --graph_type random_unlabeled_tree --graph_size 10 --n_graph 1
python3 generate_graphs.py --graph_type barabasi_albert --graph_size 10 --n_graph 1
python3 generate_graphs.py --graph_type random_unlabeled_tree --graph_size 100 --n_graph 250
python3 generate_graphs.py --graph_type random_unlabeled_tree --graph_size 250 --n_graph 250
python3 generate_graphs.py --graph_type random_unlabeled_tree --graph_size 500 --n_graph 250
python3 generate_graphs.py --graph_type random_unlabeled_tree --graph_size 1000 --n_graph 250
python3 generate_graphs.py --graph_type random_unlabeled_tree --graph_size 1500 --n_graph 250
python3 generate_graphs.py --graph_type barabasi_albert --graph_size 100 --n_graph 1
python3 generate_graphs.py --graph_type barabasi_albert --graph_size 100 --n_graph 250
python3 generate_graphs.py --graph_type barabasi_albert --graph_size 250 --n_graph 250
python3 generate_graphs.py --graph_type barabasi_albert --graph_size 500 --n_graph 250
python3 generate_graphs.py --graph_type barabasi_albert --graph_size 1000 --n_graph 250
python3 generate_graphs.py --graph_type barabasi_albert --graph_size 1500 --n_graph 250
python3 generate_graphs.py --graph_type small_world --graph_size 100 --n_graph 250
python3 generate_graphs.py --graph_type small_world --graph_size 250 --n_graph 250
python3 generate_graphs.py --graph_type small_world --graph_size 500 --n_graph 250
python3 generate_graphs.py --graph_type small_world --graph_size 1000 --n_graph 250
python3 generate_graphs.py --graph_type small_world --graph_size 1500 --n_graph 250
A in-depth analysis of the algoritm can be found here: