A search space is the set of input locations or controllable settings over which an objective function is optimized. Each tunable parameter contributes a dimension, and optimization searches combinations of parameter values for high performance. section 7.5
Airplane design
An airplane-design search space can include body length, wing span, wing angle, turbine-blade angle, and rotation speed. A simulation scores each proposed design for metrics such as speed or energy efficiency, turning parameter tuning into black-box optimization. section 7.5
Figure 7.19 The objective function of the simulated airplane design optimization problem in various two-dimensional subspaces, corresponding to pairs of tunable parameters, shown as axis labels. Bright spots indicate high objective values, which are our optimization goals; dark spots indicate low objective values.
Compare two-dimensional subspaces formed by pairs of airplane-design parameters; bright regions mark higher objective values and dark regions lower values. section 7.5
