Embodiments of the invention provide methods and apparatus for executing a multiple operand instruction. In one embodiment, a method for reward-based learning includes receiving a set of one or more exemplars, where at least two of the exemplars comprise a pair for a system, and at least one of the exemplars includes an immediate reward responsive to a pair. In another form, the optical bus comprises an optical fiber comprising an integrated array of optical beam splitters at predetermined intervals. For each of a series of frames of time, the digital signal in the time domain is converted to a digital signal in frequency domain for the frame of time. Each of the positioning boards includes a plurality of second fixing portions. The multi-layers can be interspersed with layers of Ti which can also be applied by MTCVD.