Question-and-Answer Resource for the Building Energy Modeling Community
Get started with the Help page
Ask Your Question
2

Building Energy Optimization Research Ideas

asked 2016-05-31 12:47:08 -0500

Hello,

I am fairly new to the building energy simulation field and am in a Building Systems Masters Program. I am looking for a research topic for my program thesis and am very interested in centering it around energy model optimization. One idea I had was to compare the different solutions I get from various optimization programs (GenOpt vs. Honeybee/LadyBug vs. jEplus). I would analyze the optimal solutions that each program gives and try to determine why the programs may give those solutions.

Would this be a sensible research topic? What obstacles would you imagine I would face? Any other research topics/avenues I should research instead? I am fairly open to a topic, but again, would prefer it to center around an optimization problem.

edit retag flag offensive close merge delete

Comments

1

Not sure this is the right forum for this kind of question. Can try bldg-sim@onebuilding.org listserv.

__AmirRoth__'s avatar __AmirRoth__  ( 2016-05-31 13:51:54 -0500 )edit
1

just to add, some of the algorithms (e.g. Genetic algorithms) in these software generate random solutions, so even same program will give you different results.

Waseem's avatar Waseem  ( 2016-05-31 14:03:39 -0500 )edit
1

Ladybug + honeybee don't have an optimization engine, however there are several optimization plugins for Grasshopper that can be coupled with LB+HB. The short list is Octopus, GOAT and Galapagos. As mentioned by @Waseem you will probably end up to conclude that each of the optimization engines will give you different results based on the landscape of the results and the initial population.

Mostapha Roudsari's avatar Mostapha Roudsari  ( 2016-05-31 17:07:04 -0500 )edit
1

I don't think such a comparison of optimization programs would be insightful. What solution you achieve and in what computing time is mainly a question of what optimization algorithm you use, what type of problem you ask it to solve, and how you parameterize and initialize the algorithm. Whether the algorithm is run from GenOpt, Honeybee/LadyBug or jEplus is secondary, assuming they all use parallel computing to evaluate the cost.

A comparison of optimization algorithm for single-objective building energy optimization can be found in http://dx.doi.org/doi:10.1016/j.build...

Michael Wetter's avatar Michael Wetter  ( 2016-05-31 17:55:47 -0500 )edit

Thank you for all of your comments! I will continue my search for a research topic.

Jonny K's avatar Jonny K  ( 2016-05-31 23:37:29 -0500 )edit

1 Answer

Sort by ยป oldest newest most voted
2

answered 2016-06-01 05:21:42 -0500

Yi Zhang's avatar

Hi Jonny,

As Michael and others have pointed out, simply comparing the optimisation tools such as GenOpt, LB+HB+(optimisation) and jEPlus+EA using one design problem may not give you any meaningful results. This is because they are actually designed for solving different types of problems, so you may end up comparing apples to oranges. However, bearing this in mind, there are in fact two valid research questions worth investigating: (1) a comparison on how they can be used in the design process, including the types of problems they are good at solving; and (2) a comparison of the performance of the implemented algorithms in the tools, using a set of well-defined problems whose global optima are known.

From a developer's perspective, I would love to see research addressing either of the questions, although the latter is rather harder. The Wetter/Wright paper remains so far the definitive comparison on algorithms in building optimisation. It can do with an update to include new developments in both problems and algorithms, I think.

Regards,

Yi

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Careers

Question Tools

1 follower

Stats

Asked: 2016-05-31 12:47:08 -0500

Seen: 614 times

Last updated: Jun 01 '16