INFRASTRUCTURE CAPITAL MAINTENANCE: OPTIMISE OR PRIORITISE?
Alec D Erskine, MWH UK Ltd
(free)
Some water companies are now using optimisation methods routinely to produce an optimal
investment strategy for capital maintenance. Such approaches work well for project-based
non-infrastructure work and the solvers are often based on a “branch & bound” linear
programming solvers. With infrastructure, the number of dimensions in the problem rises to
hundreds of thousands because of the number of assets involved. If options exist on the type
or timing of the intervention, the problem may become even more computationally impractical.
There are various ways to get round this. At least one commercial software package employs
a simpler optimisation approach, and makes assumptions about the family of options around
each asset and this allows it to retain the type and timing complexity.
The question arises as to the benefits of performing complex optimisations requiring massive
computational efforts given the difficulties this causes in embedding the process into a
company’s standard systems and procedures. The quality of the data also reduces the value
of highly advanced methods. A prioritisation approach may provide a near-optimal solution
because of the number of assets involved and the difference between this near-optimal
solution and a true optimal solution may not be significant. If it were not, then the advantages
of robustness, speed of reaction and ease of embedment would make such simpler
approaches very attractive to the water infrastructure industry.
KEY WORDS
Infrastructure, Optimisation, Prioritisation, Water supply, sewers, networks
