Probabilistic reasoning in intelligent systems: networks of plausible inference
Probabilistic reasoning in intelligent systems: networks of plausible inference
Dynamic Programming and Optimal Control
Dynamic Programming and Optimal Control
Multi-Objective Optimization Using Evolutionary Algorithms
Multi-Objective Optimization Using Evolutionary Algorithms
| Hi-index | 0.00 |
Microelectronics are typically critical components in a military platform, some of which may become obsolete, before the equipment life cycle end. Obsolete components may be required for a number of reasons. Components can become obsolete even before production of a platform commences. The selection of solutions for resolving obsolete components throughout a platform can be considered as a multi-criteria optimization problem where the aim is to select the most cost effective solutions for resolving a portfolio of obsolescence arisings. In this paper we consider the case where the criteria with which the options are evaluated are not point values, but probability distributions generated by a Bayesian Belief Network. We propose the use of an evaluation technique called measures of effectiveness (MOE), that can capture and use the probabilistic information associated with potential solutions. This is used with two candidate optimization techniques, Dynamic Programming (DP) and Evolutionary Algorithms (EAs), to identify cost effective solutions for resolving obsolescent components throughout a platform. Both optimization techniques were able to identify a number of solutions at different cost and MOE levels; the solutions that form the DP Pareto front dominate very slightly in places those that form the EA Pareto front.