Exercise 1.8 (Road lighting)

Consider a road divided into n segments that is illuminated by m lamps. Let pj be the power of the j th lamp. The illumination Ii of the i th segment is assumed to be j=1maijpj, where aij are known coefficients. Let Ii be the desired illumination of road i.

We are interested in choosing the lamp powers pj so that the illuminations Ii are close to the desired illuminations Ii. Provide a reasonable linear programming formulation of this problem. Note that the wording of the problem is loose and there is more than one possible formulation.

Answers

Our goal is to minimize the difference between the desired illumination [I1,,In] and the actual illumination [I1,,In]. We could choose any kind of metric to do so, but the one suitable for the linear optimization is the sum of absolute errors (cf. residual minimization from Section 1.3):

d (I,I) = i=1n |I i I i| .

Considering the fact that the powers of the lamps cannot be negative, we formulate the above optimization problem as

minimize i=1n |Ii j=1maijpj| subject topj 0,j = 1,,m .

We use the procedure from the Section 1.3 (Problems involving absolute values) to get rid of the absolute values:

minimize i=1nzi subject tozi Ii j=1maijpj i = 1,,n zi (Ii j=1maijpj)i = 1,,n pj 0 j = 1,,m

and to reformulate the above as a linear optimization problem:

minimize i=1nzi subject tozi + j=1maijpj Ii i = 1,,n zi j=1maijpj Iii = 1,,n pj 0 j = 1,,m .
User profile picture
2022-02-09 21:52
Comments