Social Business Process Simulation

Evaluate the performance of business processes one or more of their tasks are executed by a group of voluntary social network users.

Transcript

1. Jorge  Saldivar  Galli   Advisor:  Prof.  Fabio  Casa-  
2. Translate   Receipts   Translator   Can  someone   help  me  to   translate  this   receipt   from  Japanese   <receipt-­‐link>?     Facebook  post   friend   A   B   C   D   friend   Process:  Travel  Expense  Reimbursement   E   F   G   H   Task   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon   Organizations start to delegating to social network users the execution of their business process tasks [1] [1]  Fraternali  et  al.    ’12:  A  Model-­‐driven  Approach  to  Social  BPM  ApplicaGons   How  this   post  will  be   propagate?   How  their   friends  will   react?   How  will  be   the   behavior  of   the  task?  
3. Social Business Process: a business process in which one o more of its tasks are delegated to social networks users. Translate   Receipts   Translator   Can  someone   help  me  to   translate  this   receipt   from  Japanese   <receipt-­‐link>?     2.  share   friend   3.  comment  the  post   with  the  translaGon   A   B   C   D   friend   Process:  Travel  Expense  Reimbursement   E   F   G   H   Task   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
4. “ To simulate social business processes in order to evaluate their performance”! MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
5. -­‐  To  simulate  business  processes,  we  need   to  understand  the  behavior  of  their  tasks.   -­‐  The  task  behavior  is  characterize  by  a  set  of   proper<es.   -­‐  Proper<es  of  tasks  in  “tradi<onal”  business   process   ExecuGon  Time   Cost   Number  of   Resources   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
6. In  a  social  business  process,  which  are  these  proper<es?   In  social  business  process  the  organizaGon  publish  the  requirement   by  a  creaGng  a  post  in  a  social  network.   Time  spent?   People  involved?   key  to  success   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
7. Which  are  the  proper<es  that  define  the  behavior  of  tasks  executed   by  social  network  users?   How  to  define  indicators  that  evaluate  the  performance  of  processes  in   which  tasks  are  executed  by  a  voluntary  group  of  social  networks  users?   How  is  the  process  of  informa<on  propaga<on  through  social   networks?     How  social  network  users  influence  the  diffusion  of  an  informa<on?   How  social  network  users  react  regarding  a  new  informa<on  in  a  social   network?   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
8. Diffusion of Innovation! Model  of  Informa<on  Diffusion   •  Threshold  model  of  collecGve  behavior  (Granove[er  1978)     •  A  simple  model  of  global  cascades  on  random  networks  (Wa[s  2002)     •  Maximizing  the  spread  of  influence  through  a  social  network  (Kempe  et  al.  2003)     •  Network,  crowds  and  market  (Easlet  et  al.  2010)     Decision-­‐making  Process   •  Learning  From  Neighbors  (Bala  et  al.  1998)     •  Contagion  (Morris  2000)   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
9. Influential Users! Social  Network  Analysis   •  PredicGng  influenGal  users  (Ghosh  et  al.  2010)   •  Discovering  important  nodes  through  graph  entropy  the  case  of  enron  email  database   (She[y  et  al.  2005)   •  IdenGfying  influenGal  users  by  their  posGngs  in  social  networks  (Sun  et  al.  2012)     Data-­‐mining  and  Social  Network  Analysis   •  Twi[errank:  finding  topic-­‐sensiGve  influenGal  twi[erers  (Weng  et  al.  2010)     •  Everyone's  an  influencer:  quanGfying  influence  on  twi[er  (Bakshy  et  al.  2011)   •  PredicGng  influenGal  users  in  online  social  networks  (Ghosh  et  al.  2010)   •   The  influenGals:  New  approaches  for  analyzing  influence  on  twi[er  (Leavi[  et  al.  2009)   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
10. Content Popularity! Reac<on  of  users   •  Toward  predicGng  popularity  of  social  markeGng  messages  (Yu  et  al.  2011)   •  Lifespan  and  popularity  measurement  of  online  content  on  social  networks  (Sun  et  al.   2011)     Crea<on  Time   •  Time  is  on  your  side  (Mason  2012)   Topic  Analysis   •  Empirical  study  of  topic  modeling  in  twi[er  (Hong  et  al.  2010)   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
11. Social Business Process Management! Execu<on  phase   •  Business  Process  Management  with  Social  Soiware  Systems-­‐-­‐A  New  Paradigm  for  Work   OrganizaGon  (Johannesson  et  al.  2009)   •  Social  soiware  for  coordinaGon  of  collaboraGve  process  acGviGes  (Dengler  et  al.  2012)     Modeling  phase   •  A  NotaGon  for  SupporGng  Social  Business  Process  Modeling  (Brambilla  et  al.  2011)   •  Combining  BPM  and  social  soiware:  contradicGon  or  chance?  (Erol  et  al.  2010)   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
12. Post A! Post B! Property   Post  A   Post  B   #  words   68   5   age   10  hours   4  hours   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
13. Post A! Post B! Reac<ons   Post  A   Post  B   #  likes   0   4   #  comments   0   7           Why  different  reac<ons?   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
14. Task  behavior  properGes   Affect   Post  properGes   -­‐  Number  of  characters   -­‐  Topic   -­‐  Type  of  media  used   -­‐  CreaGon  Gme   -­‐  Language   Post  creator  properGes   -­‐  Number  of  neighbors   -­‐  Centrality   -­‐  Neighborhood  overlap   -­‐  Topic  of  interest   -­‐  Relevance  of  her  posts   ReacGons  generated  by   the  post   Impact  of  the  post   -­‐  PropagaGon  speed   -­‐  Number  of   different  users   interacGng  with  the   post   -­‐  Domain-­‐specific  metrics   -­‐  Number  of  interacGons   -­‐  Reached  audience   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
15. <key_indicator  name=“TranslaGonTimeAverage”>   …   <thresholds>     <range  name=“low”  lowerbound=“0”  upperbound=“30”/>     <range  name=“medium”  lowerbound=“31”  upperbound=“60”/>     <range  name=“high”  lowerbound=“61”  upperbound=“180”/>     </thresholds>   …   <funcGon  name=“AvgTranslaGonTime”  type=“SQL”>     select  avg(execuGon_Gme)     from  process_instances     where  process_name=“expense_report”   </funcGon>   …   </key_indicator>       XML   SQL   [2]  Rodriguez  et  al.    ’10:  Warehouse  model  and  diagnosGc  algorithm   Ad-hoc XML Language [2]! MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
16. Ianus  Project:  EmulaGon  Tool  for  Business  Process  FuncGonal  TesGng   BPMN   FuncGonal  TesGng     Emulator   Excel  ConfiguraGon  File   Report  of  Process  FuncGonality   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
17. •  Set  of  proper<es  that  characterize  the   behavior  of  social  business  tasks.   •  Set  of  suitable  simula<on  models  able  to   simulate  social  business  process.   •  Language  to  define  social  business  process   performance  indicators.   •  Simula<on  Tool  for  social  business  process.   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
18. •  Evaluate  the  correctness  of  the  simula<on   models  by  comparing  the  simulaGon  results   with  data  collected  from  social  network.   •  Evaluate  the  indicator  language??   •  Evaluate  the  simula<on  tool  according  to  the   evaluaGon  guideline  provides  in  [3]   [3]  Bosilj-­‐Vuksic  et  al.    ’07:  Criteria  for  evaluaGon  of  business  process  simulaGon  tool   MoGvaGon   Goal   Problem   Challenges   State  of  the   Art   Experiment   Approach   Progress   Expected   Results   ValidaGon  
19. Thanks!   QuesGons?  
20. •  T1:  Selec<on  of  a  process  class  (survey)  and   construc<on  of  the  dataset.   •  T2:  Refine  task  behavior  properGes  and   construc<on  simulaGon  models.   •  T3:  Evalua<on  and  generaliza<on  of   simulaGon  models.   •  T4:  Development  of  indicator  language.   •  T5:  Implementa<on  of  simulaGon  tool.   •  T6:  Evalua<on  of  the  approach.   •  T7:  Wri<ng  research  proposal.  
21. MoGvaGon   Goal   Challenges   State  of  the   Art   Approach   Progress   Expected   Results   ValidaGon   Organizations start to delegating to social network users the execution of their business process tasks [1] [1]  Fraternali  et  al.    ’12:  bpmn  extension  for  social  business  process  management   Evaluate   quality  of   public  metrics   Calculate   weighted   evaluaGon   results   IdenGfy   public   metrics