myxie/shadow

ScHeduling Algorithms for DAG Workflows

shadow's Issues

Add makespan plotting to visualisation

implement crowding distance

Crossover functionality

Produce json file instead of csv when generating cost matrics

check _is_feasible to determine if mutation has created an infeasible journey

Binary tournament implementation

upward_rank is not pure

The function upward_rank is not pure: it modifies the input data, a "side effect" in functional-programming lingo. Pure functions are easier to reason about, combine, test, debug and parallelize (source). This is motivation for Dask assumes functions are pure by default (docs) and "side effects are rarely used [in functional programming]" (wiki).

NSGAII* dominates function

[feature request] Allow customization of plots

Let's say I have this code:

from shadow.visualiser.visualiser import AllocationPlot

sample_allocation = AllocationPlot(solution=HEFTWorkflow.solution)
sample_allocation.plot()

And I want to customize the plot. For example, maybe I want to add vertical grid lines.

Convert daliuge graph translation from heft library into attributes json file

Edit graph attributes in the class

Generate random graphs in .json format with ccr ratio set by user

Mutation functionality