Computational-Assignment

This project consists of two main parts:

1. Econometrics Part - Statistical analysis of patient data
2. Operations Research Part - A discrete event simulator for MRI scheduling

Operations Research Part Structure

classDiagram
class Event {
+datetime start_date
+float scan_duration
+PatientType patient_type
+working_hours_till(Event)
+__lt__(Event)
}
class EventCall {
inherits Event
}
class EventScan {
inherits Event
+datetime end_date
+add_delay(float)
+calculate_overtime()
}
class PatientType {
+TYPE_1
+TYPE_2
+from_string(str)
}
class MRItype {
+TYPE_1
+TYPE_2
}
class MRI {
+MRItype type
+Set[datetime] booked_slots
+Dict[datetime,float] delays
+float slot_duration_hours
+float total_scan_time
+slot_generator(datetime)
+calculate_total_delay(datetime, float)
+store_delay(datetime, float)
+get_accumulated_delay(datetime)
+add_scan_time(duration)
+calculate_utilization(end_date, start_date)
}
class FutureEventsList {
-List[Event] heap
+add_event(Event)
+pop_event()
+peek_event()
}
class DES {
+datetime current_date
+datetime last_scheduled_date
+List[float] scan_times
+bool merged
+float total_waiting_time
+int amount_served
+float total_delay
+float total_overtime
+float max_overtime
+FutureEventsList future_list
+Dict[PatientType,MRI] MRImachines
+DefaultDict[date, List[float]] delays_by_date
+run()
+stats()
+handle_scan(EventScan)
+handle_call(EventCall)
}
Event <|-- EventCall
Event <|-- EventScan
Event --> PatientType
MRI --> MRItype
DES --> FutureEventsList
DES --> MRI
FutureEventsList --> Event

Loading

Key Components

Models

• Event (models/event.py): Base class for simulation events

  • EventCall: Patient appointment request events
  • EventScan: MRI scanning events with end time tracking and delay management
  • Handles working hours calculations and event ordering

• MRI (models/mri.py): MRI machine management

  • MRItype: Enum for machine types
  • Handles slot scheduling and availability
  • Manages delays and their accumulation throughout the day

• PatientType (models/patient.py): Patient classification

  • Enum for different patient types
  • Handles conversion from string representations

Services

• FutureEventsList (services/event_list.py): Event queue management

  • Priority queue implementation for events
  • Maintains chronological event ordering

• DES (services/simulator.py): Discrete Event Simulator

  • Main simulation logic
  • Handles event processing
  • Collects and reports statistics

Utils

• FileReader (utils/file_reader.py): Data input handling
  • Parses CSV records
  • Creates initial event list
  • Validates input data

Simulation Parameters

Operating Hours

• Working hours: 8:00 - 17:00
• Weekdays only (Monday-Friday)
• Overtime is tracked when scans extend beyond 17:00

Configuration Options

1. Input File (filePath):

   • CSV file containing patient records
   • Format: date,time,duration,patient_type
   • Example: 2024-01-01,8.5,1.2,"Type 2"

2. Scan Times (scan_times):

   • List of scan durations for each patient type
   • [Type1_duration, Type2_duration]
   • Units: hours
   • Example: [0.8, 1.2] means:
     • Type 1 patients: 0.8 hours (48 minutes)
     • Type 2 patients: 1.2 hours (72 minutes)

3. Machine Strategy (merged):

   • True: All MRI machines can handle all patient types
   • False: Dedicated machines per patient type
   • Affects how patients are assigned to machines

Running the Simulation

1. Prepare your input CSV file
2. Configure the simulation parameters
3. Run using:

python -m operations_research_part.main

Output Statistics

1. Timing Statistics

   • Average waiting time in operational hours
   • Maximum waiting time in operational hours
   • Last scheduled scan date and time
   • Last finished scan date and time

2. Efficiency Metrics

   • Total overtime used
   • Maximum overtime in a day
   • MRI utilization rates per machine type

3. Delay Analysis

   • Average delay per customer
   • Average total delay per day
   • Average delay per scan

Data Generation

The project includes scripts for generating synthetic patient data:

• econometrics part/Statistical Analysis Type 1.ipynb: Generates Type 1 patient data
• econometrics part/generate data type2.py: Generates Type 2 patient data
• econometrics part/merge_generated.py: Merges both patient types into a single dataset

Statistical Analysis

The project includes statistical analysis of:

• patient type 1
  • (econometrics part/Statistical Analysis Type 1.ipynb)
• patient type 2
  • Inter-arrival times (econometrics part/interarrivals7.py)
  • Scan durations (econometrics part/ durations4.py)

These analyses inform the parameters used in data generation and simulation configuration.