Safe Continuous-time Automated Decision Making with Mathematical Optimisation

Monash University, Australia

Updated: about 2 hours ago
Location: Melbourne, VICTORIA
Deadline: The position may have been removed or expired!

SCIPPlan  is a mathematical optimisation based automated planner for domains with i) mixed (i.e., real and/or discrete valued) state and action spaces, ii) nonlinear state transitions that are functions of time, and iii) general reward functions. SCIPPlan iteratively i) finds violated constraints (i.e., zero-crossings) by simulating the state transitions, and ii) adds the violated constraints back to its underlying optimization model, until a valid plan is found. Potential applications of this project include pandemic planning, navigation (e.g., see Figure 1 below), Heating, Ventilation and Air Conditioning control etc. The purpose of this Ph.D. project is to incorporate safety measures (e.g., with respect to uncertainty, against adversarial agents etc.) into the automated decision making of SCIPPlan.

Figure 1: Visualisation of a plan generated by SCIPPlan for an example navigation domain where the red square represents the agent, the blue rectangles represent the blocks, the gold star represents the goal location and the delta represents time. The agent can control its acceleration and the duration of its control input to modify its speed and location in order to navigate in a two-dimensional maze. The goal of the domain is to find a path for the agent with minimum makespan such that the agent reaches its the goal without colliding with the obstacle. Note that SCIPPlan does not linearise or discretise the domain to find a valid plan.

Similar Positions