/[ascend]/trunk/models/basemodel.a4l
ViewVC logotype

Contents of /trunk/models/basemodel.a4l

Parent Directory Parent Directory | Revision Log Revision Log


Revision 2651 - (show annotations) (download) (as text)
Thu Dec 13 07:29:48 2012 UTC (11 years, 6 months ago) by jpye
File MIME type: text/x-ascend
File size: 19334 byte(s)
Fixing GPL header, removing postal address (rpmlint incorrect-fsf-address)
1 (* ASCEND modelling environment
2 Copyright (C) 1998, 2007 Carnegie Mellon University
3
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 2, or (at your option)
7 any later version.
8
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
13
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>.
16 *)
17 PROVIDE "basemodel.a4l";
18 (*
19 basemodel.a4l, by Benjamin A. Allan 03/98 - Original Code
20
21 Basic definitions cmu libraries and standard methods.
22 This file is necessary for all other CMU authored ASCEND models to work
23 in ASCEND IV.
24 *)
25
26 MODEL catch_Word_model (* Bill Gates sacrificial goat *);
27 (* This MODEL does nothing.
28 * Normally catch_Word_model just gets parsed and ignored.
29 *
30 * If the user has tried to read a Microsoft Word binary file, Tcl file,
31 * or some other piece of junk as if it were an ASCEND MODEL
32 * source file, then catch_Word_model will die on an unknown
33 * syntax error.
34 * While catch_Word_model is dying the parser returns a good
35 * starting condition.
36 *
37 * Here is the message of recovery when this MODEL fails:
38 Asc-Error: Model definition "catch_Word_model" abandoned due to syntax errors.
39 Asc-Error: Rejected "catch_Word_model" at line basemodel.a4l:62.
40
41 @TODO document how this works
42 *)
43 END catch_Word_model;
44
45 (* First define the standard methods, or stand-ins which will tell
46 * us when a standard method has not been written.
47 *)
48 ADD METHODS IN DEFINITION MODEL;
49
50 METHOD ClearAll;
51
52 NOTES 'purpose' SELF {
53 This method finds anything that is a solver_var and changes
54 the .fixed flag on the var to FALSE.
55
56 This method does not change .included flags on relations
57 or return boolean, integer, or symbol variables to a
58 default value.
59 } END NOTES;
60
61 EXTERNAL asc_free_all_variables(SELF);
62 END ClearAll;
63
64 (*
65 * Geniuses make more mistakes than anyone else -- because they
66 * try more things that anyone else. Part (perhaps a very large
67 * part) of what makes a genius different from the rest of
68 * humanity is that they quickly recognize their own mistakes
69 * and move on to try something else before anyone notices
70 * they screwed up! Solving a problem as far and as fast as you
71 * can, then going back to criticize every aspect of the solution
72 * with an eye to improving it is how you usually discover right answers.
73 *
74 * The authors of ASCEND (geniuses or not we'll
75 * leave to our users to decide) have found that it is
76 * best to do things such as writing mathematical MODELs and
77 * writing mathematical modeling software in ways which
78 * makes our mistakes (or your mistakes) very easy to detect.
79 *
80 * Below we describe a methodology (pun intended) which can
81 * help make anyone who can solve a quadratic equation
82 * a mathematical modeling expert. This methodology helps
83 * you to avoid screwing up and to find out about it when you have.
84 *
85 * The ASCEND system will not force you to write standard
86 * methods in your models. :-( METHODs of the sort we advocate
87 * here make your MODELs much easier to use and
88 * much more reliable. They pay off in the short run as well
89 * as the long run. These are _guidelines_, not _laws_: real
90 * genius requires knowing when to color outside the lines. :-)
91 *
92 * If you do not write the standard methods, your MODEL will
93 * inherit the ones given here. The "ClearAll" and "reset"
94 * methods here will work for you if you followed the guidelines.
95 * The other methods contain STOP statements which will warn you
96 * that you have skipped something important, should you accidentally
97 * end up calling one of them.
98 *
99 * The following methods should be redefined by each
100 * reusable library MODEL that REFINES this root MODEL.
101 * Models that do not supply proper versions of these
102 * (and possibly other) methods are very hard to reuse.
103 *
104 * The premise of this method design is that we can
105 * write the _self methods incrementally, building on the
106 * already tested methods of previous MODEL parts we are
107 * reusing. In this way we never have to write a single huge method
108 * that directly checks 100s of variables in a hierarchy.
109 *
110 * The _all methods are methods which simply "top off" the
111 * _self methods. With an _all method, you can treat
112 * just a part of a larger simulation already built
113 * as a self-contained simulation.
114 *
115 *)
116
117 (*
118 * Usually discovery of the information you need to write the methods
119 * proceeds in the order that they appear below:
120 * check, default, specify, bound, scale.
121 *)
122
123 METHOD check_self;
124 NOTES 'purpose' SELF {
125 This method should be written first, though it is run
126 last. Just like they taught you in elementary school,
127 always check your work. Start by defining criteria for a
128 successful solution that will not be included in the
129 equations solved and then computing those in this method.
130 As you develop your MODEL, you should expect to revise the
131 check method from time to time, if you are learning
132 anything about the MODEL. We frequently change our
133 definition of success.
134
135 When a mathematical MODEL is solved, the assumptions that
136 went into writing (deriving) the equations should be
137 checked. Usually there are redundant equations available
138 (more than one way to state the physical MODEL
139 mathematically). These should be used to check the
140 particularly tricky bits of the MODEL.
141
142 Check that the physical or intuitive (qualitative)
143 relationships among variables ch you expect to hold are
144 TRUE, especially if you have not written such relationships
145 in terms of inequalities in the MODEL equations.
146
147 In some models, checking the variable values against
148 absolute physical limits (temperature > 0{K} and
149 temperature < Tcritical for example) may be all that is
150 necessary or possible. Do not check variable values against
151 their .lower_bound or .upper_bound, as any decent algebraic
152 solver or modeling system will do this for you.
153
154 If a check fails, use a STOP or ERROR statement to notify
155 yourself (or you MODEL using customer) that the solution
156 may be bogus.
157
158 Currently only STOP is implemented.
159 STOP raises an error signal and issues an error message;
160 STOP normally also stops further execution of the method
161 and returns control to a higher level, though there are
162 interactive tools to force method execution to continue.
163 STOP does not crash the ASCEND system.
164
165 } END NOTES;
166
167 (* STOP {Error! Standard method "check_self" called but not written in MODEL.}; *)
168
169 END check_self;
170
171 METHOD check_all;
172
173 NOTES 'purpose' SELF {
174 When solving only a part of a simulation, it is necessary to check
175 the models and variables passed into the part as well as the
176 locally defined parts and variables. This method should check
177 all the received models and variables, then check the local
178 stuff.
179 } END NOTES;
180
181 (* STOP {Error! Standard method "check_all" called but not written in MODEL.}; *)
182 RUN check_self; (* intentionally _second_ *)
183
184 END check_all;
185
186 METHOD defaults;
187 (*
188 * This is a kluge for interfaces that still think of
189 * 'defaults' as the standard method.
190 *)
191 RUN default_self;
192 STOP {GUI (or somebody) called non-standard method defaults. Call forwarded to default_self before stopping here.};
193 END defaults;
194
195 METHOD on_load;
196 NOTES 'purpose' SELF {
197 This method adds improved ability to perform stuff when a model is first loaded.
198 By default, just 'default_self' will be run (this was the previous behaviour).
199 Any model that has an on_load method can override this behaviour however.
200 Note that this behaviour applies only in the C++/python interface at this stage.
201 } END NOTES;
202 RUN default_all;
203 END on_load;
204
205 METHOD default;
206 NOTES 'purpose' SELF {
207 This method exists for the purpose ofOVERRIDING atom defaults in the local
208 MODEL. Hopefully this approach can replace the current practise of writing
209 'default_self' methods for most MODELs. It should be superior, since using
210 'default' instead of 'default_self' will always result in ALL variables in a
211 model being reset to default values, rather than only those explicitly stated
212 by the modeller.
213 } END NOTES;
214 (*STOP {it works!};*)
215 (* nothing here *)
216 END default;
217
218 METHOD default_self;
219 NOTES 'purpose' SELF {
220 This method should set default values for any variables
221 declared locally (IS_A) to the MODEL. It should run
222 default_self on _all_ the models that are declared locally
223 (with IS_A) in the MODEL also. If the atoms you use to
224 define your variables have a suitable default already, then
225 you do not need to assign them a default in this method.
226
227 This method should not run any methods on MODEL parts that
228 come via WILL_BE in the definition's parameter list. This
229 method also should not change the values of variables that
230 are passed in through the parameter list.
231
232 Sometimes there will be nothing for this method to do.
233 Define it anyway, leaving it empty.
234
235 When a top-level simulation is built by the compiler, this
236 method will be run at the end of compilation by the
237 compiler. See notes in on_load method for new behaviour in
238 the PyGTK GUI.
239 } END NOTES;
240 EXTERNAL defaultself_visit_childatoms(SELF);
241 EXTERNAL defaultself_visit_submodels(SELF); (* overwrite ATOM defaults explicit nested code if needed *)
242 RUN default; (* local overrides *)
243 END default_self;
244
245 METHOD default_all;
246 NOTES 'purpose' SELF {
247 This method assumes that the arguments to the MODEL
248 instance have not been properly initialized, as is
249 frequently the case in one-off modeling efforts. This
250 method should run the default_self method on each of the
251 parts received through the parameter list and should give
252 appropriate default values to any variables received
253 through the parameter list. After these have been done, it
254 should then call default_self to take care of all locally
255 declared default needs.
256 } END NOTES;
257 (* INITIALISATION OF PARAMETERS IS NOT IMPLEMENTED YET *)
258 RUN default_self;
259 END default_all;
260
261 METHOD specify;
262 NOTES 'purpose' SELF {
263 * Assuming ClearAll has been run on the MODEL, this method
264 * should get the MODEL to a condition called 'square':
265 * the case where there are as many variables with .fixed == FALSE
266 * as there equations available to compute them.
267 * This is one of the hardest tasks ever invented by mathematicians
268 * if you go about it in the wrong way. We think we know the right way.
269 *
270 * Actually, 'square' is a bit trickier to achieve
271 * than simply counting equations and variables.
272 * Solver, such as QRSlv in ASCEND, may help greatly with the bookkeeping.
273 *
274 The general approach is to:
275
276 (1) Run "specify" for all the parts (both passed in and locally defined)
277 that are not passed on into other parts.
278
279 (2) Fix up (by tweaking .fixed flags on variables) any difficulties
280 that arise when parts compete to calculate the same variable.
281
282 (3) Use the remaining new local variables to take care of any leftover
283 equations among the parts and any new equations written locally.
284
285 At all steps 1-3
286 Pay special attention to indexed variables used in
287 indexed equations; frequently you must fix or free N or
288 N-1 variables of a set sized N, if there are N matching equations.
289 In general, if you think you have specify correctly written, change
290 the sizes of all the sets in your MODEL by one and then by two
291 members. If your specify method still works, you are using sets
292 correctly.
293
294 When writing models that combine parts which do not share
295 very well, or which both try to compute the same variable
296 in different ways, it may even be necessary to write a WHEN
297 statement to selectively TURN OFF the conflicting equations
298 or MODEL fragments. An object or equation USEd in a WHEN
299 statement is turned off by default and becomes a part of
300 the solved MODEL only when the conditions of some CASE
301 which refers to that object are matched.
302
303 The setting of boolean, integer, and symbol variables which
304 are controlling conditions of WHEN and SWITCH statements
305 should be taken care of in the specify method.
306
307 There is no 'one perfect "specify"' for all purposes. This
308 routine should merely define a reasonably useful base
309 configuration of the MODEL.
310
311 Other specify_whatElseYouWant methods can (should) also be
312 written.
313
314 The name of a method is a communication tool. Please use
315 meaningful names as long as necessary to tell what the
316 method does. Avoid cryptic abbreviations and hyper-
317 specialized jargon known only to you and your three friends
318 when you are naming methods; however, do not shy away from
319 technical terms common to the engineering domain in which
320 you are modeling.
321
322 } END NOTES;
323
324 (* STOP {Error! Standard method "specify" called but not written in MODEL.}; *)
325
326 END specify;
327
328 METHOD reset;
329 NOTES 'purpose' SELF {
330 This method gets the MODEL to some standard starting state,
331 though not necessarily the most useful starting state for a
332 particular application. In Chem. Eng. terms, this method
333 establishes a base case.
334
335 There is no 'one perfect "reset"' for all purposes. This
336 routine should merely define a reasonably useful base
337 configuration of the MODEL.
338
339 Other reset_whatElseYouWant methods can (should) also be
340 written.
341
342 Normally you do not need to write this method: your models
343 will inherit this one unless you override it (redefine it)
344 in your MODEL.
345 }
346 END NOTES;
347
348 RUN ClearAll;
349 RUN specify;
350
351 END reset;
352
353 METHOD values;
354 END values;
355
356 METHOD bound_self;
357 NOTES 'purpose' SELF {
358 Much of the art of nonlinear physical modeling is in
359 bounding the solution.
360
361 This method should update the bounds on _locally_ defined
362 (IS_A) variables and IS_A defined MODEL parts. Updating
363 bounds requires some care. For example, the bounds on
364 fractions frequently don't need updating.
365
366 A common formula for updating bounds is to define a region
367 around the current value of the variable. A linear region
368 size formula, as an example, would be:
369
370 v.upper_bound := v + boundwidth * v.nominal;
371 v.lower_bound := v - boundwidth * v.nominal;
372
373 Care must be taken that such a formula does not move the
374 bounds (particularly lower bounds) out so far as to allow
375 non-physical solutions. Logarithmic bounding regions are
376 also simple to calculate.
377
378 Here boundwidth IS_A bound_width;
379 boundwidth is a real variable (but not a solver_var) or a
380 value you can use to determine how much "wiggle-room" you
381 want to give a solver. Small powers of 4 and 10 are usually
382 good values of boundwidth.
383
384 Too small a boundwidth can cut off the portion of number
385 space where the solution is found. Too large a bound width
386 can allow solvers to wander for great distances in
387 uninteresting regions of the number space.
388
389 This method should not bound variables passed into the
390 MODEL definition or parts passed into the definition.
391 } END NOTES;
392
393 (* STOP {Error! Standard method "bound_self" called but not written in MODEL.}; *)
394 END bound_self;
395
396 METHOD bound_all;
397 NOTES 'purpose' SELF {
398 This method should be like bound_self except that it bounds the
399 passed in variables and calls bound_self on the passed in parts.
400 It should then call bound_self.
401 } END NOTES;
402
403 (* STOP {Error! Standard method "bound_all" called but not written in MODEL.}; *)
404 RUN bound_self;
405 END bound_all;
406
407 METHOD scale_self;
408 NOTES 'purpose' SELF {
409 Most nonlinear (and many linear) models cannot be solved without
410 proper scaling of the variables.
411
412 This method should reset the .nominal value on every real
413 variable in need of scaling. It should then call the
414 scale_self method on all the locally defined (IS_A) parts
415 of the MODEL. 0.0 is the worst possible nominal value. A
416 proper nominal is one such that you expect at the solution
417 the quantity
418
419 abs(variable/(variable.nominal))
420
421 to be around 1 (in the range of [0.1..10] or [0.01..100]).
422
423 Variables (like fractions) bounded such that they cannot be
424 too far away from 1.0 in magnitude probably don't need scaling
425 most of the time if they are also bounded away from 0.0.
426
427 Some solvers, but not all, will attempt to scale the
428 equations and variables by heuristic matrix-based methods.
429 This works, but inconsistently; user-defined scaling is
430 generaly much superior.
431
432 ASCEND makes it easy to do. You scale the variables, which
433 can only be done well by knowing something about where the
434 solution is going to be found (by being an engineer, for
435 example.) Then ASCEND can calculate an appropriate
436 equation-scaling by efficient symbolic methods.
437
438 This method should not change the scaling of models and
439 variables that are received through the parameter list of
440 the MODEL.
441 } END NOTES;
442
443 (* STOP {Error! Standard method "scale_self" called but not written in MODEL.}; *)
444 END scale_self;
445
446 METHOD scale_all;
447 NOTES 'purpose' SELF {
448 This method should be like scale_self above except that it also
449 should scale the variables and models received through the
450 parameter list. It should then call scale_self to take care of
451 the local variables and models.
452 } END NOTES;
453
454 (* STOP {Error! Standard method "scale_all" called but not written in MODEL.}; *)
455 RUN scale_self;
456 END scale_all;
457 END METHODS;
458
459 MODEL cmumodel();
460 NOTES
461 'purpose' SELF {
462 This MODEL does nothing except provide a root
463 for a collection of loosely related models.
464 If it happens to reveal a few bugs in the software,
465 and perhaps masks others, well, what me worry?
466 BAA, 8/97.
467 }
468 'methods' SELF {
469 This MODEL also provides a hook to put in candidates for
470 becoming ascBuiltin global methods. Global methods may be
471 overridden by local definitions.
472 BAA, 3/98.
473 }
474 END NOTES;
475
476 END cmumodel;
477
478 MODEL testcmumodel();
479 (*
480 * All CMU test models, of whatever sort should ultimately be
481 * rooted here or be a final refinement of a reusable MODEL.
482 *)
483 METHODS
484 METHOD values;
485 (*
486 * In a final application MODEL, you should record at least one set of
487 * input values (values of the fixed variables and guesses of key
488 * solved-for variables) that leads to a good solution.
489 * Do this so noone need reinvent that set the next time
490 * you use the MODEL or someone picks the MODEL up after you.
491 *)
492 (* STOP {Error! Standard method "values" called but not written in MODEL.}; *)
493 END values;
494
495 METHOD specify;
496 (* STOP {Error! Standard method "specify" called but not written in test MODEL.}; *)
497 END specify;
498
499 METHOD ClearAll;
500 EXTERNAL asc_free_all_variables(SELF);
501 END ClearAll;
502
503 METHOD reset;
504 (* This method gets the MODEL to some standard starting state,
505 * though not necessarily the most useful starting state for
506 * a particular application. In Chem. Eng. terms, this method
507 * establishes a base case.
508 * There is no 'one perfect "reset"' for all purposes. This
509 * routine should merely define a reasonably useful base configuration
510 * of the MODEL.
511 * Other reset_whatElseYouWant methods can (should) also be
512 * written.
513 *
514 * Normally you do not need to write this method: your models
515 * will inherit this one unless you override it (redefine it)
516 * in your MODEL.
517 *)
518 RUN ClearAll;
519 RUN specify;
520 END reset;
521 END testcmumodel;
522
523 MODEL your_site_models();
524 (* if you create a library to share with the net which is
525 * not just an end application of Carnegie Mellon models,
526 * please create an empy root MODEL such as this and use
527 * it as the origin of your library in the same way that
528 * we use cmumodel as the origin of our libraries.
529 * Thank you.
530 *)
531 END your_site_models;

john.pye@anu.edu.au
ViewVC Help
Powered by ViewVC 1.1.22