This is the code for the paper "Adaptive Partially-Observed Sequential Change Detection and Isolation". The paper is accepted in Technometrics and the Arxiv version is provided here.

Code

1. srpabstract.py: A class that do SRP statistics

• Initialization:
  • p: Number of dimensions
  • c: scale vector, Target meanshift is c * M
  • k: Number of failuer Mode
  • M: Failure Mode Mean Matrix of k failure modes: p * k
  • nsensors: number of selected sensors
  • Ks: Number of selected failure mode
  • L: control limit, set to -1 if not initialized yet.

You need to override the following functions

• compute_log_LRT: Compute the log liklihood ratio,
• compute_index: Compute the index function to decide the best sensing allocation

Other implemented method:

• compute_monitoring_statistics: Computed SRP statistics using the Top-R rules
• compute_monitor_batch: Compute the monitoring results for batch of samples

2. TSSRP.py: Implement original TSSRP method

• srpabstract initialization except

  • mode: which testing statistics to use, Default to T2

• compute_index: Sensor index becomes the failure mode index

3. ExtendedTSSRP.py

• srpabstract initialization except

  • mode: which testing statistics to use, Default to T2

• compute_index

  • Mode T2 Default, using summation of log SRP.
  • Mode T1 , using summation of SRP, no closed forms and using greedy algorithm
  • Mode T1_Max , using summation of SRP, no closed formsusing Max approximation

4. spc.py: A generic class for process monitoring using simulation

• Initialization:

  • monitor_statistics: given input of n_batch * Tmax * dimensions, return the monitoring statistics of size n_batch * Tmax denote
  • data_gen_func0: Generate normal samples, return n_batch * Tmax * dimensions
  • data_gen_func1: Generate abnormal samples, return n_batch * Tmax * dimensions

• phase1:

  • Generate iterations (= number of seeds) of each data with n_batch size
  • Use Binary search to find the control limit L where the ARL0 is fixed

• phase2:

  • Generate iterations (= number of seeds) of each data with n_batch size
  • Return ARL1

Dataset

There are three datasets that used in the paper:

1. The dataset is offered in temperature.mat. Thermal imaging monitoring of 3D printing: The dataset used is publically available at figshare Scenario 3. data
2. Tonnage dataset, which is offered in tonnage.mat.
3. COVID-19 monitoring of different counties in WA. The dataset is derived from the JHU. The WA dataset used in the paper has been also stored in Infection rate.