6 Chapter 1. Introduction OR (a) Bike system (b) Bike components (c) (Partial) Fault Tree of a bike Bike cannot ride safely Chain breaks Front wheel fails Back wheel fails Wheels fail AND OR Cassette fails Disc brake rotor fails Figure 1.4: Example of a Fault Tree model of a bike. probability density functions, the model can generate quantitative metrics that support decision-making for e!ective system management. One of the main challenges associated with FTs is building the model itself. To address this, in this dissertation, we used Multi-Objective Evolutionary Algorithms, which we discuss in the next section. Multi-Objective Evolutionary Algorithms Set of solutions Non-dominated solutions Market cost Modern computer Pareto-front Old computer Feasible objective space Performance Figure 1.5: Computer market cost vs performance (example). Multi-Objective Evolutionary Algorithms (MOEAs) are population-based search strategies with conflicting objectives to be simultaneously optimised in a multidimensional space (Deb, 2011). To explain their concept, let us consider the following example: assume that older computers have lower performance due to outdated technology and lower market costs, while modern computers have higher performance and higher market costs. Plotting market cost versus performance for a set of computers creates a visualisation like Figure 1.5.
RkJQdWJsaXNoZXIy MjY0ODMw