Hierarchical planning and multi-level scheduling for simulation-based probabilistic risk assessment

Abstract

Simulation of dynamic complex systems-specifically, those comprised of large numbers of components with stochastic behaviors-for the purpose of probabilistic risk assessment faces challenges in every aspect of the problem. Scenario generation confronts many impediments, one being the problem of handling the large number of scenarios without compromising completeness. Probability estimation and consequence determination processes must also be performed under real world constraints on time and resources. In the approach outlined in this paper, hierarchical planning is utilized to generate a relatively small but complete group of risk scenarios to represent the unsafe behaviors of the system. Multi-level scheduling makes the probability estimation and consequence determination processes more efficient and affordable. The scenario generation and scheduling processes both benefit from an updating process that takes place after a number of simulation runs by fine-tuning the scheduler's level adjustment parameters and refining the planner's high level system model.

Supervised dynamic probabilistic risk assessment: Review and comparison of methods

“…A more specific example is [16], where simulation (and therefore a DDET) is guided through hierarchical planning in the SimPRA framework [51]. The level of detail of the physical models is also adjusted when "level control nodes" are reached in the DDET to help reduce the computational time -i.e., the level control nodes adaptively adjust the level of detail in the simulation.…”

“…We assume the simulator has a constant computational time, meaning it will always take the same amount of time to reach a branching point. According to [16], SimPRA performs a depth-first search in a simulation tree, first looking at the most relevant scenarios as defined by a biasing function (i.e., heuristic) -however, all possible branches of the tree will be reached eventually. As the branches are generated iteratively by the scheduler module and, according to the worst-case condition 1, this algorithm can be classified as an Iterative deepening depth-first search (IDDFS) algorithm 2 .…”

“…In the average-case performance for [16], end-states will likely be found in levels before the maximum tree depth, biasing will lead to the most relevant scenarios first, and the level of detail will be adjusted from level control nodes, reducing the simulation's computing time. Thus, the average-case time complexity here is likely linear:…”

Automatic Generation of Failure Mechanism Propagation Scenario via Guided Simulation and Intelligent Algorithm

“…Although this hybrid system includes hardware, software, and human elements, different failure modes, dynamic features, and the physical characteristics of the system form the basis of the simulation model [9]. Nejad [10] extended the guided simulation method from the binary state to the multi-state and incorporated hierarchical planning and multi-level scheduling into the simulation [11]. The author provided a case study of a Lunar Reconnaissance Orbiter satellite system.…”