Wednesday, September 9, 2009

Why not?

Moral: Wherein we ask 'why not? which question can go a couple of ways. We're going to look at one of these. Actually, there is a third that is the mantra of the toddler learning about its will, that of others, and boundaries (say, good behavior), but we'll not go there either.


The query could indicate an awareness that something ought to be done or to happen, say like realizing one suggestion from the President's speech. Stay in school and be a good student (as good as you can be). Or, why not build a new plane with entirely new processes and material, that is, re-position the decision framework on an unknown center plus relax constraints along about all decision axes? In other words, the little engine that could. Or, why not pursue globalization? That's not the focus, for now.

The query could relate to reasons for something not happening, such as causal analysis tries to effect. Oh yes, risk analysis tries to do this before the fact. That use is what we're going after in order to pull together some 'un' words with a coherent message. For instance, 'undecidability' has be thrown about here, thanks to its use by the Vienna School. And, 'underdetermined' has been used, mainly due to its usefulness in resolving how we think about several dilemmas. We need to add 'unpredictable' as the risk people don't seem to know the concept.

Let's try this sequence of using these concepts (U U U N, unfortunately, not as meaningful as TANSTAAFL) in order to explain why these have appeared as well as to start discussion about what their applicability means to our difficult situations. There are other orders; this one seems to have some power for explication.
  • Unpredictable - Actually, this ought to be used for the fact that we have no a priori knowledge of how events will unfold, for the most part. Of course, some stable events recur without our involvement or worry. For instance, days succeed one another. Yet, how many lost their wealth, and income, of late due to some idiocy (not necessarily their own) that can be explained? Do we know of any problematic programs? Say, has weather forecasting in your area been 100% on (whatever that would mean) this past summer? That the quants have run off after their stochastic taming attempts we'll be looking at further. That CAE folks have broken the rules of map/territory differentiation is another variation of mis-handling of that which this sequence denotes.
  • Underdetermined - Well, what does this mean? We can use the wiki definition. Since predictability, and its complement (ah, is that so?), are figured using mathematics, albeit a subset dealing with things called statistics, let's use that type of story here. Basically, if we have a series of equations and unknowns to be determined, we like them to match in cardinality. That is, if there are too few equations, then we'll have to sort out the extra unknowns by various schemes (assumptions, swags, etc.). If there are too many equations (too few unknowns), then we're on the other side of the bad boat. The case, with computational modeling, is just this. Too much fudging (sorry, techie folks, but the truth has to be told). Also, note that we talked about unknowns. The equations would contain variables, to boot, whose definition and handling can be problematic.
  • Undecidable - Now, we have to mention Turing's, and related, work, briefly. Then, we can just discuss the issues involved with the complicated concept, in general. It is more than just a computation problem, as we see from Hilbert's query. To shorten a potential long story, we can say that some issues are not related to time, sequence, or any type of order. Yet, we can always re-frame things to be decidable, within limits (meaning choice, allowable error, etc.), of course.
  • NP (Nondeterministic polynomial) -- We're using this to introduce complexity. You see, even re-framing does not guarantee some solution unless we're exceptionally good (which some people are) or lucky (again, we see this). Yet, it's the computational aspect that is troublesome, such as using approximation, heuristics, and more. Why? We have had many of years to learn about human dynamics; yet, there continues to be intractable problems. Adding in the computer just exacerbates things incredibly. But, for thinking of complexity in terms of using the computer for solutions take the chess game. Well, that little game, on a small board, is highly complex. Something as seemingly simple as checking if two circuits are equivalent is, to boot.
Having gone through this, can you see other orders? It's the chicken/egg thing. For instance, 'undecidable' might be made the basis, with incompleteness being a key thing. The order, shown in the list, comes about since we're dealing with things that have gone awry due to our immaturity with using computational resources.

One could suggest a decision framework based upon this order that is operationally effective (with a bow of acknowledgment to quasi-empiricism). By the way, with NP at the core, how can we proceed? Ah, that is the essence of truth engineering's contribution.

But, this question of 'why not?' and UUUN ought to be part of the program management vernacular, and jargon. So, we'll continue this next time.

Note: An important concept will be separability. On predictability, some say, well, the sun will rise. Sure. But, weather was added in for this reason: will it be clear enough to see the rise? You see, NASA faces this problem all the time with shuttle launches. So, various windows of opportunity can be identified, with some ranking of certainty. Yet, the operational scheme is to be ready to take advantage; and, how many times does weather cause a mission delay? Now, if the parts of engineering that relate to business would just go back to their true framework, perhaps some of the hubris would diminish. The issues of underdetermination, undecidability and NP are important just because predictability, as a highly mathematicized approach, can be easily mis-used. Too, people play power games in these complex situations where someone who fails to solve an essentially intractable problem has his head removed. Sheesh. How the hell can we expect to learn without failure? Those who try and fail ought to be honored. Except in this situation that is very much apropos for mention and current: playing silly games with other peoples' monies in a risky fashion is verboten, by definition. Because of near-zero, we ought not even honor those who win. Stupid game, folks.


01/15/2015 -- This week, this post is getting read. Great! Nice little piece of work (kidding, in part) so many years ago. ... At last, a series that will establish the basis and extensions, as required. We are going to go back to some simple and come forward to the modern, complicated economy. Why? My long chain of ancestors (inherited via Prof. Lucio Arteaga) is one motivation.

03/11/2012 -- We need to update this, especially with a nod to Alan M. Turing.

10/07/2010 -- Several principles need to be explored, such as the ergodic one.

12/08/2009 -- Consider current CEOs in relation to Paul. Not fair? Well, these guys/gals have set themselves upon some supposed plane that is above the rest of us.

11/29/2009 -- Rationality and risk. Need a new look.

11/08/2009 -- The gigantic chimera needs proper attention.

10/16/2009 -- 201K <-- 401K --> 25601K, this denotes the current financial gaming.

Modified: 01/15/2015

No comments:

Post a Comment