cov_time_trend

pyoptex.doe.cost_optimal.cov.cov_time_trend(time=1, cost_index=0)

Covariance function to account for time trends. The cost at cost_index is assumed to represent some form of time. Every time, the level of the time trend is increased by one unit.

For example, if time is 2 and we have runs with cumulative time [0, 1, 2, 3, 4, 5], the added time column will be [-1, -1, 0, 0, 1, 1].

Parameters

timefloat

Every time in cumulative cost, the time column progresses one equidistant step.

cost_indexint

The index in the multi-cost objective to look at.

Returns

covfunc(Y, X, Zs, Vinv, costs)

The covariance function.