Institute of Computer Science

MTAT.03.231 – Business Process Management

Regular Exam – 27 May 2014

Notes:
• The exam is open-book, open-laptop, open-Internet. You are not allowed to
share information with anyone during the exam other than the lecturer.
• The exam should be submitted electronically using the submission button on
the course web page.

Task  1.  Process  Automation  (35  points)  

We consider the following loan application assessment process at a bank (herein
called the loan provider). The process starts when a customer (the applicant) submits
a loan application. Whenever a loan application is submitted, a loan officer checks it
for completeness and consistency. If the application is incomplete, it is returned to the
applicant together with comments from the loan officer. The applicant has two weeks
to re-submit the application; otherwise the loan application is closed.
Once a loan application is deemed to be complete, it undergoes two assessments in
parallel: (i) a loan risk assessment performed by a financial officer, and (ii) an
appraisal of the property for which the loan was requested. The appraisal is performed
by a property appraiser at the bank. As a result of the risk assessment, the application
is assigned a risk score, which is an integer between 1 and 100. The output of the
property appraisal is an estimation of the market value of the house, as determined by
the appraiser.
Once the risk assessment and the property appraisal have both been performed, the
loan officer assesses the applicant’s eligibility (preferably the same loan officer who
checked the application initially). If the applicant is not eligible, the application is
rejected. In case of an application for an amount exceeding 50 000, a second loan
officer (different from the one who declared the application eligible in the first place)
must confirm the eligibility. Once an application is confirmed to be eligible
(i.e. accepted), a loan officer (any loan officer) prepares a loan agreement. The loan
agreement must then be printed and signed by the applicant and submitted to the loan
provider either via file upload.
The signed application and repayment agreement must be received within 14 days
after acceptance of the loan application. If the signed documents are submitted on
time, a loan officer marks the loan application as “granted” and the process ends.
Otherwise, the application is marked as “declined” and the process ends.
A loan application contains the following data fields:
• Submission date (to be filled in automatically by the system)
 • Applicant data: name, surname and date of birth
• Contact information: e-mail and cell phone number
• Current address: Street name and street number, city and postal code
• Employment details: name of employer and monthly net revenue
• Property information:
o Address: street name and street number, city and post code
o Purchasing price
• Loan information: Amount requested, number of years, start date
Other attributes/relationships may be added if required by the process.
The loan agreement associated to a loan application contains the following fields:
• Borrower data: name, surname and date of birth (same as applicant’s details)
• Loan amount – determined by the loan officer who prepares the agreement
• Start date – determined by loan officer
• Duration of the loan (number of years) – determine by loan officer
Task. Implement the above process as an executable process model using Bizagi. For
this task, you must submit a dump of the Bizagi project database (.bak file). As a
precaution, retain a working copy the process in your laptop until the exam is graded.

Task  2.  Process  Mining  (15  points)  

The event log attached to this exam (see the Assessment web page) corresponds to a
process for handling reviews of articles (papers) submitted for publication to a
journal. The event log consists of 100 cases (papers). Each paper is sent to three
different reviewers. The reviewers are invited to write a report. However, reviewers
often do not respond. As a result, it is not always possible to make a decision after a
first round of reviewing. If there are not enough reports, then additional reviewers are
invited. This process is repeated until a final decision can be made (accept or reject).
Starting from this log, answer the following questions using ProM 5.2:
a) Using three different discovery algorithms derive three Petri nets from L and
using conformance checking, determine which is the best one (the Petri net
that has the highest fitness with respect to L). We call this Petri net P. Design a
BPMN model from P. Include in the model the routing probabilities associated
to each decision point.
b) What are the bottlenecks in the process execution?
c) What is the average waiting time needed to start the acceptance decision task
after that the reviews have been collected?
For this task, submit the three Petri nets, the BPMN model and textual answers to
questions (a), (b) and (c). In the textual answer to question (a), you should indicate
which process mining algorithms you used to generate the Petri nets, and which
algorithm produced the “best” result.
You can submit a single Word, ODT or PDF file for his task, or if needed, you can
submit the Petri nets and BPMN diagram as separate files. A single file is preferred.