Technical Report: Agile Systems Engineering: Kanban Scheduling


Report Number:  Technical Report SERC-2015-TR-111

Report Name:  Agile Systems Engineering: Kanban Scheduling

Publication Date: December 22, 2015

Project: Agile-Lean Software Engineering Evaluating Kanban In SE


Technical Report RT-111: Agile Systems Engineering: Kanban Scheduling

Background

Developing,  creating or  evolving systems  of systems  (SOSs)  present significant  systems engineering and management problems. Dahmann and Baldwin characterize these problems as stakeholder involvement,  governance,  operational focus,  acquisition,  test and evaluation,  boundaries and interfaces, and performance and behavior [1]. All systems face some of these problems,  but the  uniqueness of  the dynamics  and resulting  communication issues  in a  SoS require a significant ability for adaptation within the system development community, as well as among the stakeholders. The principles for addressing these issues are no different from those required for  any good  systems engineering  and development  activity  [2].  Implementation of  those principles in SoS environments, however, is a much thornier problem.

Agile and lean philosophies  have shown  to be  effective in  supporting adaptation  within development  and evolution    [3],  [4],  [5].  Complicated,  large systems  of systems  in rapid  or continuous deployment  environments,  where requirements  are not  precise and  can change  or emerge quickly, find traditional approaches inadequate.

In 2011, the Systems Engineering Research Center began to investigate alternative management and governance  approaches for  these complex  environments,  including a  concept for  an integrated  multi-­‐level  network of  pull scheduling  systems based  on explicit,  transparent and continuously updated value of work  [6], [7], [8].  This Kanban-­‐based Scheduling System Network (KSSN) concept was developed based on the following capabilities:

  • Coordinate multiple levels of development activity across multiple system components with diverse and possibly disjoint or isolated development groups cover the entire program lifecycle, particularly through Test and  Evaluation.
  • Support analysis and decision making at every level
  • Flexibly schedule work considering value across the system of systems
  • Balance work in progress (WIP) across resources with SoS organizational capacity to improve flow
  • Make visible to all levels progress toward capability development and deployment
  • Establish a basis for continuous improvement in a rapidly changing environment

Difficulties  in  validating  this  concept  in  vivo  led  to  the  decision  to  create  a  broad  simulation  environment  that  would  allow  in  vitro  experimentation  with  KSSN,  but  also  be  applicable  to studying  other  mechanisms,  singly  and  in  concert,  operating  in  a  range  of  organizational  structures (including all four types of systems of systems identified in [1]) and handling different kinds, durations, complexity, and volumes of work flow. We believe that establishing statistically significant evidence across various combinations of mechanisms, organizations and work flows, as well as providing  a  suitable simulation  “sandbox”  for adopters  to perform  their own  experiments will provide a level of confidence that in vivo experimentation (piloting) is low risk and provides value to adopters.

Additional Researchers:

  • Dr. Jeffrey Smith, Auburn University
  • Dr. Levent Yilmaz, Auburn University
  • Mr. Donghuang Li, Auburn University
  • Mr. Saicharan Reddy Chada, Auburn University
  • Mr. Alexey Tregubov, University of Southern California

Researchers

Collaborators