Tuesday, December 16, 2008

Target weight

Body Mass Index is a generally accepted way to look at how your height and weight should relate to each other. BMI is the ratio of your weight to the square of your height

An ideal BMI for an adult male is someplace in the range of about 20.7-28.4.

Using the upper end of that range you can estimate a target weight (including a factor for conversion between feet/pounds and metric measurements) as

W = .0404 * H2

where H is your height in inches.

For me, at 68 inches, that gives a target weight of about 186 pounds.

Ideally I think I should probably weigh in less than that, but it provides a good target at the upper range of acceptable weight for me.

Labels: , , ,

Tuesday, December 2, 2008

Goal programming

For now I'm going to stick with the linear programming model of the diet problem that I've been talking about

minimize carbohydrates
subject to constraints on
protein
calorie intake
calories from fat
sodium
etc, etc

But there is an alternative modeling technique (which I might try later) called goal programming.

In goal programming I'd want to minimize deviations from goals where the goals are various intake levels of
carbohydrates
protein
calories
calories from fat
sodium

I'm already doing that to some extent in my linear programming model by introducing slack variables to allow for exceeding or falling short of various levels of nutrients and mathematically treating these slack variables as "foods" with very high carb counts. That's just a mathematical modeling trick to allow the solution algorithm to work even in situations where the constraints are at odds with each other and a feasible solution that meets all the constraints can't be found.

So changing the model to a goal programming model rather than a linear programming model isn't going to be a big leap, but I'm not going to do it for a couple of weeks. I'll be thinking about it though.

Labels: ,