cov_double_time_trend

pyoptex.doe.cost_optimal.cov.cov_double_time_trend(time_outer=1, time_inner=1, cost_index=0)

Covariance function to account for double time trends. The inner time column is reset every time the outer time column resets. The cost at cost_index is assumed to represent some form of time. Every time_outer, the level of the outer time trend is increased by one unit. Every time_inner, the level of the inner time trend is increased by one unit, being reset whenever the outer time trend changes.

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

Parameters

time_outerfloat

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

time_innerfloat

Every time_inner in cumulative cost, the second time column progresses one equidistant step. It is reset whenever time_outer is adjusted.

cost_indexint

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

Returns

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

The covariance function.