Predictive control applied to managing corporate production and transport capacities

Introduction:

By applying simple dynamical models for product quantities in production plants, warehouses, and stores, students will design an algorithm which uses predictions of daily demand sequences in stores to decide on the production quantities and transportation quantities from production plants to warehouses and from warehouses to stores, for every day, one month in advance. It is also necessary to ensure minimal production and transportation costs for the given predicted demand. Analysis of the daily operation of the algorithm over long term will also be conducted. To solve the resulting linear program, professional software package for mathematical programming IBM CPLEX will be used.

 

Tasks:

  1. To get acquainted with the linear programming interface in IBM CPLEX

  1. To model the production-transportation network in nodal form with production plants, warehouses, and stores acting as linear dynamical systems    

  1. To formulate the problem as a linear program using the above mentioned model and solve it with one of the IBM CPLEX’s LP solvers

  2. To extend the designed algorithm with a graphical user interface to input the necessary data 

  3. To extend the designed algorithm with further problem specifications like taking demand uncertainty into account

  4. To extend the designed algorithm with production and transportation costs quantization (introducing additional shifts in a production plant or additional trucks to a transport line)

Additional information:

  • Number of students: 1
  • For undergraduate or graduate students