1. Introduction to Systems Thinking Dr Mike Yearworth

Reader in Systems Engineering

©  University  of  Bristol  2014  

What are Systems?

•  “A  system  is  a  set  of  parts  which,  when  combined,  have  qualiAes  that  are  not  present  in  any  of  the  parts  themselves.  Those  qualiAes  are  the  emergent  properAes  of  the  system.”  

•  “We  are  increasingly  concerned  with  complex  systems,  in  which  the  parts  interact  with  each  other  and  with  the  outside  world  in  many  ways”  

•  “The  relaAonships  between  the  parts  determine  how  the  system  behaves.”  

•  “IntuiAon  rarely  predicts  the  behaviour  of  complex  systems.”  

2

(EllioK  &  Deasley,  2007)  

2  

What is Systems Thinking?

•  “Viewing  situaAons  holisAcally,  as  opposed  to  reducAonisAcally,  as  a  set  of  diverse  interacAng  elements  within  an  environment  

•  Recognising  that  the  relaAonships  or  interacAons  between  elements  are  more  important  than  the  elements  themselves  in  determining  the  behaviour  of  the  system  

•  Recognising  a  hierarchy  of  levels  of  systems  and  the  consequent  ideas  of  properAes  emerging  at  different  levels,  and  mutual  causality  both  within  and  between  levels  

•  AccepAng,  especially  in  social  systems,  that  people  will  act  in  accordance  with  differing  purposes  or  raAonaliAes”  

3

(Mingers  &  White,  2010)  

3  

What is Systems Thinking?

•  “a  discipline  for  seeing  wholes.  It  is  a  framework  for  seeing  interrelaAonships  rather  than  things,  for  seeing  paKerns  of  change  rather  than  staAc  snapshots…systems  thinking  is  a  discipline  for  seeing  the  ‘structures’  that  underlie  complex  situaAons,  and  for  discerning  high  and  low  leverage  change”  

•  “…an  epistemology  which,  when  applied  to  human  acAvity  is  based  upon  the  four  basic  ideas:  emergence,  hierarchy,  communicaAon,  and  control  as  characterisAcs  of  systems.  When  applied  to  natural  or  designed  systems  the  crucial  characterisAc  is  the  emergent  properAes  of  the  whole.”  

4

(Senge,  1990)  

(Checkland  &  Scholes,  1999)  

4  

Hard  systems  tradi,on   So0  systems  tradi,on  Assumes  the  world  contains  systems  that  can  be  engineered  

Assumes  the  world  is  problemaAcal  but  can  be  explored  with  systems  models  

Assumes  systems  models  to  be  models  of  the  world  (ontologies)  

Assumes  systems  models  to  be  intellectual,  conceptual  constructs  (epistemologies)  

Modelling  is  oriented  to  goal  seeking,  opAmisaAon,  and  predicAon  

Modelling  is  oriented  to  learning,  exploraAon,  and  commitment  

Talks  the  language  of  “soluAons”   Talks  the  language  of  “issues”  

Advantages   Advantages  Allows  use  of  powerful  quanAtaAve  techniques  (simulaAon,  visualisaAon…)  

Is  open  to  all  stakeholders  and  keeps  in  touch  with  human  interests  

Disadvantages   Disadvantages  May  loose  touch  with  the  actual  problem  situaAon;  ownership  and  control  issues  

Does  not  produce  final  answers,  accepts  that  inquiry  is  never  ending  

The “Hard” and “Soft” systems traditions compared

5  

Adapted  from  (Checkland  &  Holwell,  2004)  

Systems paradigms

6

(Pidd, 2004; Kotiadis & Mingers, 2006)

So#$$

Hard$$

So#$$

Hard$$

Hard$$ So#$$

858 Journal of the Operational Research Society Vol. 57, No. 7

a b c dHARD OR - data analysis

HARD - data mining HARD SOFT

T HARDSOFTOR Time Nz -SSM

SOFT Hard and soft intertwined.

Figure 1 The soft and hard paradigm taken from Pidd (2004a) and Brown et al (2006).

simple resistance to multiparadigm thinking' (p 533) for example to abandon all but one paradigm as suggested by Pfeffer (1993). Secondly, in many cases it 'provides an overall framework that mixes and combines terms and implications of arguments grounded in different paradig- matic assumptions without considering the relationship between the assumptions themselves' (p 533). We will focus on paradigm crossing as it provides a theoretical basis and explanation of multiparadigm work in practice.

Paradigm crossing is about how multiple paradigms are dealt with by an individual researcher without ignoring them, as the integrationist position, or refusing to confront them, as the incommensurability position. Schultz and Hatch (1996) focus on strategies for paradigm crossing and they add a new one to the existing ones in the literature. The existing strategies are the sequential, the parallel, the bridging, and interplay (Schultz and Hatch's strategy). In the sequential strategy, the relationship between paradigms is linear and movement from one paradigm to another is unidirectional. In the parallel strategy, different paradigms are all applied on an equal basis instead of sequentially. The sequential and parallel strategies leave the boundaries of each paradigm deployed intact but in the bridging strategy the boundaries are more permeable. In fact, Gioia and Pitre (1990), who first articulated the bridging strategy, argued that paradigm boundaries are ill defined and blurred, and that it is 'difficult, if not impossible, to establish exactly where one paradigm leaves off and another begins' (p 592). Therefore, the boundaries are better conceived as transition zones.

The interplay strategy 'refers to the simultaneous recogni- tion of both contrasts and connections between paradigms rather than differences' (p 534). Researchers using this strategy will transpose the findings from one paradigm in such a way that they inform the research conducted in a different paradigm. Therefore, the researcher can move back and forth between paradigms allowing cross-fertilization between the paradigms while maintaining diversity. The main difference of the interplay with the other strategies is 'in the nature of the relationship it constructs between the researcher and the multiple paradigms it specifies' (p 535). It is this approach that is closest to that advocated by Mingers (1997b) and appears to fit Pidd's (2004a) (Figure 1) last

illustration, which is essentially a representation of a paradigm strategy.

Tashakkori and Teddlie (1998) also explore the paradigm strategies for combining qualitative and quantitative meth- odologies and categorise the sequential and parallel under mixed method studies but they do not refer to the bridging or interplay strategies although they acknowledge a mixed model studies category that could incorporate them.

This debate about strategies is an attempt to understand and communicate the multimethodology work undertaken that could be classified as multiparadigm. This debate is in its infancy and it is interesting to see the authors that are currently engaged attempting to communicate with the audience not only by words but also with pictures as if words cannot go deep enough under the surface into the subconscious where most of the work arguably took place. A picture is worth a thousand words and perhaps in such situations a more useful communication tool.

Cultural difficulties

Peoples' assumptions about the world and how to deal with its problems is to some extent a cultural issue that has resulted through socialization and education. In manage- ment science, there are communities particularly educational ones that are perceived as more hard OR focused or to have a more balanced number of hard and soft OR specialists than others. However, the latter are thought to be fewer particularly as many working in the field of management science emerge from a variety of positivist disciplines, for example, mathematics, computer science, engineering, etc. Undoubtedly this must affect both the type of research or projects undertaken in these departments as well as the student's experience and attitude to problem solving immediately after university. It is logical to assume that this culture also feeds into industry. PSMs are probably not even considered by the majority of management scientists in the first instance when considering a problematic situation and many will simply turn to the old familiar approaches. In the event that they are considered for use either alone or with a traditional hard OR method there is probably a degree of fear about being competent in their use.

On the other hand, Pidd (2004a) makes some interesting points about the clients view about someone being able to

6  

Why Systems? Messes, swamps and wicked problems…

•  Aim  is  interven*on  in  a  problem  situaAon,  not  knowledge  gathering  for  its  own  sake  

•  No  definiAve  formulaAon  of  the  problem  situaAon  •  There  is  no  stopping  rule,  the  problem  situaAon  is  on-­‐going  •  IntervenAons  are  not  right  or  wrong,  there  is  no  immediate/ulAmate  

test  of  an  intervenAon,  but  only  viewed  as  good/bad  •  IntervenAons  are  ‘one-­‐shot’,  no  trial-­‐and-­‐error  (experiments),  every  

intervenAon  counts  significantly,  they  are  essenAally  unique  •  No  enumerable,  exhausAvely  describable,  set  of  intervenAons  •  Problem  situaAons  can  be  considered  as  symptoms  of  other  problems  •  IntervenAons  can  be  contested  at  the  level  of  explanaAon,  there  is  

likely  to  be  conflicAng  evidence/data  

7

Adapted from (Rittel & Webber 1973)

7  

Why Systems? Messes, swamps and wicked problems…

•  Problem  situaAons  involving  many  interested  parAes  with  different  perspecAves  (worldviews)  

•  Problem  situaAons  that  are  not  well  defined  •  There  is  difficulty  in  agreeing  objecAves  of  intervenAons    •  Success  of  intervenAons  requires  creaAng  agreement  

amongst  parAes  involved  •  There  are  many  uncertainAes  and  lack  of  reliable  (or  any)  data  •  Almost  certainly  trying  to  work  across  the  boundary  between  

human  acAvity  and  the  artefacts  of  engineering  

Adapted from (Mingers 2011)

8  8  

Taking Action: Problem Structuring Methods (PSMs)

•  Methods, not mathematical, but structured and rigorous and based on qualitative, diagrammatic modelling

•  Allow for a range of distinctive views to be expressed/explored/accommodated and allow for multiple and conflicting objectives

•  Encourage active participation of stakeholders in the modelling process, through facilitated workshops and cognitive accessibility

•  Can facilitate negotiating a joint agenda and ownership of implications of action

•  Significant uncertainty is expected and tolerated •  Operate iteratively •  Aim is for exploration, learning, and commitment from

stakeholders

9

Adapted from (Mingers, 2011; Mingers & Rosenhead, 2004; Rosenhead, 1996)

9  

Example PSM: Soft Systems Methodology (SSM)

10

Adapted by John Davis from (Checkland & Scholes 1999)

10  

Generic properties of PSMs – summary

1. AcAon  to  improve    2. Systemic  approach    3. AdaptaAon  and  creaAvity  4. Methodological  lessons  5. Worldviews  6. Wicked/messy  problem  situaAon  7. InteracAve,  iteraAve  and  therapeuAc  8. SubjecAvity    9. Limits  

11  

Adapted from (Yearworth & White 201x)

(Yearworth & White, 201x) Theoretical underpinnings – see handout Aspect     Defini,on  

Improvement  AcAvity   A  structured  way  of  approaching  systemic  intervenAon  has  been  taken,  which  was  designed  to  lead  to  improvements  in  a  problemaAc  real-­‐world  situaAon  through  a  set  of  purposeful  acAviAes  

Systemic  Approach   The  problem  structuring  approach  used  systems  ideas  (including  boundary,  hierarchy,  communicaAon  and  control),  which  i)  are  appropriate  to  context,  ii)  theoreAcally  adequate,  and  iii)  supported  by  appropriate  systems  modelling  

AdaptaAon/  CreaAvity  

Conscious  thought  and  creaAvity  must  have  gone  into  how  the  problem  structuring  approach  was  adapted  or  elements  combined  for  the  parAcular  problem  situaAon  

Methodological  Lessons   Use  of  the    problem  structuring  approach  yielded  methodological  lessons,  extracted  by  conscious  reflecAon  

Worldviews   The  process  of  problemaAsaAon  recognised  that  problems  are  construct  of  an  individual’s  mind,  they  do  not  exist  independently  of  human  thought.  These  constructs  are  defined  by  an  individual’s  “worldview”,  the  problem  structuring  approach  acknowledged  these  and  worked  with  them  

Messiness   The  problem  context  in  which  the  problem  structuring  approach  was  used  was  recognised  as  messy|wicked|swampy  following  definiAons  such  as  contained  in  (Ackoff,  1979,  1981;  RiKel  &  Webber,  1973;  Rosenhead,  1992;  Vennix,  1999).  

InteracAve/  IteraAve/  TherapeuAc  

The  intervenAon  in  the  problem  situaAon  has  come  about  through  sharing  of  “percep*ons,  persuasion  and  debate”  in  a  parAcipaAve  group  sekng  using  an  interacAve  and  iteraAve  approach.  The  facilitator  or  owner  of  the  problem  structuring  approach  adopted  a  stance  that  was  “interac*ve/therapeu*c,  not  expert”  

SubjecAvity   In  the  approach  taken  it  has  been  recognised  that  the  stakeholders  of  the  problem  situaAon  are  not  “divorced  from  the  problem”  and  that  they  could  not  act  as  objecAve  “outsiders”  as  in  the  ‘hard’  systems  tradiAon  

Limits     Approaches  to  problem  structuring  might  unwikngly  suffer  from  inter  alia  bounded  raAonality,  inadequacy  of  organizaAonal  language  to  supply  adequate  conceptual  terminology,  applicaAon  of  non-­‐appropriate  methodology,  “spurious  saliency”  etc.  The  approach  used  demonstrated  that  it  dealt  with  such  conceptual  limitaAons  including  building  experAse  in  the  use  of  problem  structuring  methods  

12  

Exercise – Restating the problem situation

•  In  groups  •  Quick  5  minute  brainstorm  to  

idenAfy    –  Key  issues  –  Stakeholders  involved  –  Assump*ons  you  are  making  as  you  approach  the  “definiAon  of  integrated  and  comprehensive  urban  plans  addressing  the  efficient  energy  flows  across  various  sectors,  focusing  on  the  long  term”  

Systems!Thinking!for!comprehensive!city!Efficient!Energy!Planning! December!2011!Acronym:!STEEP! !

! Page!No.!4!of!114!

know[how!for!strategic!integrated!planning,!involvement!of!all!stakeholders!with!responsibility!for!action!during!the!implementation!process!and!the!availability!of!tools!and!instrument!for!decision!making,!planning!and!monitoring.!!Until! very! recently,! a! large! number! of! cities! have! undertaken! different! actions! aimed! to! promote! a! more!sustainable!use!and!production!of! energy.!Generally,! these!actions! focus!on!sector!specific!problems!and!are!often!undertaken!in!isolationc!they!do!not!take!into!account!all!the!energy!flows!of!the!city!(e.g.!energy,!transport,!water,! digital! infrastructure,! etc.).! Furthermore,! most! of! the! time! the! actions! taken! are! not! integrated! into! a!strategic! approach! to! urban! planning,! which! integrates! city! land[use! planning! alongside! other! city! planning!processes!e.g.!economic!development!strategies,!sustainable!development!plans,!digital!masterplanning,!etc.!!It!seems!evident,!however,! that! the!definition of integrated and comprehensive urban plans!addressing! the!efficient! energy! flows! across! various! sectors,! focusing! on! the! long term,! achieves! much! better results and bigger impact than individual actions,!since!the!complementarities!and!synergies!between!actions!are!taken!into!consideration.!This project will focus on creating integrated Smart City Plans.!!

!

!FIGURE 1. INTEGRATED SMART CITY PLAN

13  

References Ackoff,  R.  L.  (1979).  Future  of  OperaAonal-­‐Research  is  Past.  Journal  of  the  Opera*onal  Research  Society,  30(2),  93-­‐104.  doi:  

10.2307/3009290  Ackoff,  R.  L.  (1981).  The  Art  And  Science  Of  Mess  Management.  Interfaces,  11(1),  20-­‐26.    Checkland,  P.,  &  Holwell,  S.  (2004).  "Classic"  OR  and  "soq"  OR  -­‐  an  asymmetric  complementarity.  In  M.  Pidd  (Ed.),  Systems  

Modelling:  Theory  and  Prac*ce.  Chichester:  John  Wiley  &  Sons,  Ltd.  Checkland,  P.,  &  Scholes,  J.  (1999).  SoM  Systems  Methodology  in  Ac*on:  Including  a  30-­‐year  retrospec*ve.  Chichester:  Wiley.  EllioK,  C.,  &  Deasley,  P.  (2007).  Crea*ng  Systems  That  Work:  Principles  of  engineering  systems  for  the  21st  century  Retrieved  from    

 hKp://www.raeng.org.uk/educaAon/vps/pdf/rae_systems_report.pdf    KoAadis,  K.,  &  Mingers,  J.  (2006).  Combining  PSMs  with  hard  OR  methods:  the  philosophical  and  pracAcal  challenges.  Journal  of  

the  Opera*onal  Research  Society,  57(7),  856-­‐867.  doi:  10.1057/palgrave.jors.2602147  Mingers,  J.  (2011).  Soq  OR  comes  of  age-­‐but  not  everywhere!  Omega-­‐Interna*onal  Journal  of  Management  Science,  39(6),  

729-­‐741.  doi:  10.1016/j.omega.2011.01.005  Mingers,  J.,  &  Rosenhead,  J.  (2004).  Problem  structuring  methods  in  acAon.  European  Journal  of  Opera*onal  Research,  152(3),  

530-­‐554.  doi:  10.1016/s0377-­‐2217(03)00056-­‐0  Mingers,  J.,  &  White,  L.  (2010).  A  review  of  the  recent  contribuAon  of  systems  thinking  to  operaAonal  research  and  management  

science.  European  Journal  of  Opera*onal  Research,  207(3),  1147-­‐1161.  doi:  10.1016/j.ejor.2009.12.019  Pidd,  M.  (2004).  Complementarity  in  systems  modelling.  In  M.  Pidd  (Ed.),  Systems  Modelling:  Theory  and  Prac*ce.  Chichester:  

John  Wiley  &  Sons  Ltd.  Yearworth,  M.,  &  White,  L.  (201x).  The  non-­‐codified  use  of  Problem  Structuring  Methods  and  the  need  for  a  generic  consAtuAve  

definiAon.  European  Journal  of  Opera*onal  Research  -­‐  In  Review.      

14  

References RiKel,  H.  W.  J.,  &  Webber,  M.  M.  (1973).  Dilemmas  in  a  General  Theory  of  Planning.  Policy  

Sciences,  4(2),  155-­‐169.    Rosenhead,  J.  (1992).  Into  the  Swamp  -­‐  The  Analysis  of  Social-­‐Issues.  Journal  of  the  Opera*onal  

Research  Society,  43(4),  293-­‐305.  doi:  10.1057/jors.1992.44  Rosenhead,  J.  (1996).  What's  the  problem?  An  introducAon  to  problem  structuring  methods.  

Interfaces,  26(6),  117-­‐131.  doi:  10.1287/inte.26.6.117  Senge,  P.  M.  (1990).  The  FiMh  Discipline:  The  Art  and  Prac*ce  of  the  Learning  Organiza*on.  

London:  Random  House.  Vennix,  J.  (1999).  Group  model-­‐building:  tackling  messy  problems.  System  Dynamics  Review,  

15(4),  379-­‐401.    Yearworth,  M.,  &  White,  L.  (201x).  The  Non-­‐Codified  Use  of  Problem  Structuring  Methods  and  

the  Need  for  a  Generic  ConsAtuAve  DefiniAon.  European  Journal  of  Opera*onal  Research  -­‐  In  Review.    

15