Lead time and cycle time agile optimisation simulator

Introduction

This repository attempts to find team configuration, work size, team member capacity and backlog order that delivers work at the lowest lead time possible. This repository comes with experiments for Scrum, Scrum with less handovers, Kanban and Waterfall.

To run, download, load in Visual Studio Code, under Run and Debug press "Run Simulation". Click on debug console, you will greeted with the following output for all experiments:

##############################START###################################
###  Small Team Test Experiment

...

# Assumptions
1: Lead Time does NOT follow normal distribution (Nonparametric) => true
2: Cycle Time does NOT follow normal distribution (Nonparametric) => true
3: Two random Lead Time control experiments come from same distribution (Null-Hypothesis is true) => true
4: Team member idle does NOT follow normal distribution (Nonparametric) => true
 

# Control 
  Total mean man-days: original 18.6, actual 18.7
Conditions: 
  Capacity :  PO => Capacity 0.3 Dev => Capacity 1
## Lead Time 
  Days Deviation: 9.8
*When* delivered: 
  First 25% delivered on day 4.75, 50% 8.25, 75% 12, last 25% 31.75
## Cycle Time
  *Time taken* to deliver once started: 
  25% has taken 3 day(s), 50% 5.25, 75% 8.25, last 25% 18.25
## Constraint
  Member: Dev, Idle Deviation: 3.1
## Team Members
 Dev => idle days 1, turn count: waiting 24, preq 3, feedback 0
 PO => idle days 3, turn count: waiting 0, preq 18, feedback 0


# Experiment 
  Total mean man-days: original 23.1, actual 23.7
Conditions: 
  Sort : OrderByLargest
  Capacity : PO => 1, Dev => 3.9
## Lead Time 
  Days Deviation: 11.9
*When* delivered: 
  First 25% delivered on day 0.25, 50% 1.25, 75% 2.75, last 25% 8.75
## Cycle Time
  *Time taken* to deliver once started: 
  25% has taken 0.25 day(s), 50% 0.75, 75% 1.75, last 25% 5
## Constraint
  Member: Dev, Idle Deviation: 0.9
## Team Members
 Dev => idle days 2, turn count: waiting 2, preq 0, feedback 6
 PO => idle days 2, turn count: waiting 0, preq 1, feedback 0


# Control vs Experiment (Null Hypothesis): Significant difference (Rejected)
##############################END###################################

Optimisation methods

Branch and Bound optimisation is used for backlog sort. Polynomial regression with Powell's method is used for team capacity prediction.

Customisation

It is possible to setup your own experiments by extending SoftwareTest class like so:

export class YourTest extends SoftwareTest {

    public readonly Name: string = "YourTest";

    public readonly Description: string = `Your own simulation`;

    public readonly teamConfig = new TeamConfig([
            new MemberConfig("Product Owner", 10/37, 8/10, 4/100),
            new MemberConfig("UX", 10/37, 4/10, 10/100),
            new MemberConfig("Architecture", 5/37, 5/10, 5/100),
            new MemberConfig("Back-End", 37/37, 8/10, 30/100),
            new MemberConfig("Front-End", 37/37, 8/10, 30/100),
            new MemberConfig("Test", 37/37, 10/10, 20/100),
            new MemberConfig("Product Owner Sign Off", 1/37, 10/10, 1/100)],
        [
                [0, 1/3, 1/10, 1/5,   1/5,   1/5,   0],
                [1, 0,   1/10,   0,   1/3,   1/5,   1/20],
                [1, 1,      0, 1/5,  1/10,    0,   0],
                [1, 0,      1,   0,   1/2,  1/2,   1/20],
                [1, 1,      1,   0,     0,  1/2,   1/20],
                [1, 1,      0,   1,     1,    0,   1/10],
                [0, 0,      0,   0,     0,    1,  0],
        ]
    );
}

Then it can be registered in the main.ts:

const experiments = new Array<Test>(new SmallTeamTest(), new ScrumTest(), new KanbanTest(), new ScrumPartialStackTest(), new WaterfallExperiment(), new YourTest());

If you made it this far...

This simulator was built to test assumptions about knowledge work / software development lead time, if you are interested in this consider checking out below.

Research - Rejuvenating Agile operations by putting lead and cycle time front and centre

Notebook - Exploration of lead time dynamics in Sprint scenario

Notebook - Lead Time minimisation through team capacity

Article - Lead Time Driven Delivery - Part 0 - Introduction

Article - Lead Time Driven Delivery - Part 1 - Learning to see

Article - Lead Time Driven Delivery - Part 2 - Learning from data

Article - Lead Time Driven Delivery - Part 3 - Focus on results, not methods

Article - Lead Time Driven Delivery - Part 4 - Stabilise through embedded testing

Article -Lead Time Driven Delivery - Part 5 - Practical and closing thoughts

I find this really interesting, if you have any questions about this repository please feel free to contact me.