I have collected a sizeable amount of papers on radical polymerization kinetics. i also know that there exist excellent commercial programs such as e.g. Predici, but they are very expensive. Studying the papers I came to the conclusion that the method of moments is a relatively reasonable method which can be systematically applied to various copolymerization, even to bulk polymerization problems with apparently less variables and computational power requirements than e.g Monte-Carlo based methods. Is it worth to try write an own program (for non-commercial purposes) based on such, relatively simple set of equations. What are the pitfalls? My other problem is that I do not want to learn more complicated programming languages. I used to prepare even relatively complicated algorithms using Basic language - but I hardly found now simple Basic compilers that can be run in the Windows environment and which allow the transfer of the data to e.g. Excel for plotting. Sorry for this outmoded approach, but I would appreciate your comments.
What you describe is a relatively tough task. Not so much solving the equations, as the input/output requirements. The normal approach would be to solve the equations in a language like C, or Fortan, which allow for efficient handling of large arrays of numbers. Then, either wrap a shell around this so it will work in a Windows environment, or, alternatively, use two separate programs: one to solve the equations, another to convert your output. Personally, I prefer the latter approach as it grants greater flexibility.
Unfortunately, this requires two separate skill-sets: 1) to write an efficient and stable numerical solver for the equations. This is something that people working in some field of numerical physics/chemistry/mathematics should be able to do. 2) to write a user interface, or data conversion routine in a Windows environment. This is a task more suited to people with a computer-science background.
To learn either, or both skill-sets to write a single program seems a lot of work. Would it be possible to hire one or more persons to do the job, or perhaps have it done at a local university as part of a thesis project? There are many universities that offer this kind of service. Obviously not for free, but it may work out cheaper than buying commercial software, which, indeed, is very expensive.
A final consideration would be the support. This is a definitive advantage of commercial software. A privately written code is very difficult to maintain over long periods of time, because they often rely on one individual. This is also a downside to having the code written at a university. There is no guarantee that any kind of support, let alone further improvements, will be possible, once the initial project is finished.
Dear Allard, thank you very much for your expert opinion and practical advice. I will try to find subcontractors locally. Thanks Gyorgy
Dear Pavel, my main motivation to try my own way is that I can adjust fuly the porgram to my needs and it can be developed step by step. Firts the interfaton system for relatively simple systems, as homopolymerization, later taking into account gel effect, non-isothermal pehnomena and finally copoymerizations of various kinds. Not being a mathematician I used to solve differential equations by self-adatptive step length, so to say by brute force methods. It is possible that this does not work with a set a differential equations having very different rate constants. The advantage of the method of moments seems to be the fact, that a relatively low number of variable and equations are used and can be extended to various situations. I will try with simple cases and if I fail, I give up. Thanks for your advice anyway. Gyorgy
- Badges
- Science topic