File: linearalgebra.mac

package info (click to toggle)
maxima 5.47.0-9
  • links: PTS
  • area: main
  • in suites: forky, sid
  • size: 193,104 kB
  • sloc: lisp: 434,678; fortran: 14,665; tcl: 10,990; sh: 4,577; makefile: 2,763; ansic: 447; java: 328; python: 262; perl: 201; xml: 60; awk: 28; sed: 15; javascript: 2
file content (315 lines) | stat: -rw-r--r-- 12,218 bytes parent folder | download | duplicates (4)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
/*
   Author Barton Willis
   University of Nebraska at Kearney
   Copyright (C) 2005, Barton Willis

   Brief Description: Maxima code for finding the nullspace and column space 
   of a matrix, along with code that triangularizes matrices using the method
   described in ``Eigenvalues by row operations,'' by Barton Willis.

   This is free software  you can redistribute it and/or
   modify it under the terms of the GNU General Public License,
   http://www.gnu.org/copyleft/gpl.html.

   This software has NO WARRANTY, not even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
*/

put('linearalgebra,1,'version);

/* The translator generates code that doesn't work when 
   translate_fast_arrays is true. So set translate_fast_arrays
   to false.
*/

eval_when([translate, compile], translate_fast_arrays : false);

eval_when(translate,
	  declare_translated(ptriangularize_with_proviso,locate_matrix_entry,hankel,toeplitz,
			     nullspace,require_integer,columnspace,rowswap,rowop,require_symbol,
	                     mat_fullunblocker,mat_trace,mat_unblocker, column_reduce,good_pivot,
                             hipow_gzero, mat_norm))$

eval_when([batch,load,translate,compile],
  put('infolevel, 'silent, 'linearalgebra),    
  load("load-linearalgebra-lisp-files"));
  
require_integer(i, pos, fn) := 
  if integerp(i) then true else 
     error("The", pos, "argument to" ,fn, "must be an integer");

require_symbol(x, pos, fn) :=
  if symbolp(x) or subvarp(x) then true else
     error("The", pos, "argument to", fn, "must be a symbol");

request_rational_matrix(m, pos, fn) :=
  if every('identity, map(lambda([s], every('ratnump,s)), args(m))) then true else
    print("Some entries in the matrix are not rational numbers. The result might be wrong.");

dotproduct(a,b) := block([scalarmatrixp : true],
  require_matrix(a,"first", "dotproduct"),
  require_matrix(b,"second", "dotproduct"),
  if matrix_size(a) # matrix_size(b) or second(matrix_size(a)) # 1 then 
     error("Arguments to dotproduct must be vectors with the same size"),
  ctranspose(a) . b);

nullspace(m) := block([nr, nc, acc : set(), proviso : true, pv, prederror : false],
  
  require_unblockedmatrix(m, "first", "nullspace"),
  /* nc and nr are the sizes of the transpose of m */
   
   [nc, nr] : matrix_size(m),
   m : triangularize(addcol(transpose(m), ident(nr))),
   for row : 1 thru nr do (
     pv : locate_matrix_entry(m, row, 1, row, nc, lambda([s], compare(s,0) # "="), 'bool),
     
     if not(listp(pv)) then (
       acc : adjoin(transpose(genmatrix(lambda([ii,j], m[row,j+nc]),1,nr)),acc))
     else (
       proviso : proviso and notequal(m[row, second(pv)],0))),
   
   if proviso # true then (	
     print("Proviso: ",proviso),	
     put(concat(outchar, linenum), proviso, 'proviso)),
   
   subst('span, 'set, subset(acc, lambda([s], some(lambda([x], compare(x,0) # "="), s)))));


nullity(m) := length(nullspace(m));

orthogonal_complement([v]) :=  block([sz],
  require_unblockedmatrix(first(v)," ","orthogonal_complement"),
  sz : matrix_size(first(v)),
  if second(sz) # 1 then error("Each argument must be a column vector"),
  for vi in v do (
    require_matrix(vi," ","orthogonal_complement"),
    if matrix_size(vi) # sz then error("Each vector must have the same dimension")),
  nullspace(transpose(rreduce('addcol, v))));

locate_matrix_entry(m, r1, c1, r2, c2, fn, rel) := block([im, cm, mf, e, 
     nr, nc, ok : true, frel],
  require_unblockedmatrix(m, "first", "locate_matrix_entry"),
  require_integer(r1, "second", "locate_matrix_entry"),
  require_integer(r2, "second", "locate_matrix_entry"),
  require_integer(c1, "third", "locate_matrix_entry"),
  require_integer(r2, "fourth", "locate_matrix_entry"),
  nr : length(m),
  nc : length(first(m)),
 
  if (c1 > c2) or (r1 > r2) or (r1 < 1) or (c1 < 1) or (r2 < 1) or (c2 < 1) or 
     (c1 > nc) or (c2 > nc) or (r1 > nr) or (r2 > nr) then error("Bogus submatrix"),

  mf : if rel = 'min then 'inf else if rel = 'max then 'minf else 0,	
  frel : if rel = 'max then ">" else if  rel = 'min then "<" else if rel = 'bool 
    then lambda([x,y],x) else error("Last argument must be 'max' or 'min'"),
  im : 0,
  cm : 0,
  for i : r1 thru r2 while ok do (
    for j : c1 thru c2 while ok do (
       e : apply(fn, [m[i,j]]),
       if is(apply(frel, [e, mf])) then (
         im : i,
         cm : j,
         if rel = 'bool then ok : false,
         mf : e))),
  if im > 0 and cm > 0 then [im, cm] else false);
         
columnspace(a) := block([m, nc, nr, cs : set(), pos, x, proviso : true, algebraic : true, c_min],
  require_unblockedmatrix(a, "first","columnspace"),
  [nr, nc] : matrix_size(a),
  if nc = 0 or nr = 0 then (
    error("The argument to 'columnspace' must be a matrix with one or more rows and columns")),  

  m : triangularize(a),
  c_min : 1,
  for i : 1 thru nr while c_min <= nc do (
   if listp(pos : locate_matrix_entry(m, i, c_min, i, nc, lambda([x], x # 0), 'bool)) then (
     c_min : second(pos) + 1,
     if constantp(x : part(part(m, first(pos)),second(pos))) = false then (
       proviso : proviso and notequal(ratsimp(x), 0)),
     cs : adjoin(col(a,second(pos)) ,cs))),
  
  if proviso # true then (	
    print("Proviso: ",proviso),	
    put(concat(outchar, linenum), proviso, 'proviso)),
  funmake('span, listify(cs)));

/* Maxima's rank function doesn't complain with symbolic matrices. 
I don't approve of rank(matrix([a,b],[c,d])) --> 2. This version will
print the proviso warnings.
*/

linalg_rank(m) := length(columnspace(m));

rowswap(m,i,j) := block([n, p, r],
  local(p),
  require_matrix(m, "first", "rowswap"),
  require_integer(i, "second", "rowswap"),
  require_integer(j, "third", "rowswap"),
  n : length(m),
  if (i < 1) or (i > n) or (j > n) then error("Array index out of bounds"),
  p : copymatrix(m),
  r : p[i],
  p[i] : p[j],
  p[j] : r,
  p);

columnswap(m,i,j) := transpose(rowswap(transpose(m),i,j));

/*  row(m,i) <-- row(m,i) - theta * row(m,i) */

rowop(m,i,j,theta) := block([p : copymatrix(m), listarith : true],
  local(p),
  p[i] : p[i] - theta * p[j],
  p);

/*  col(m,i) <-- col(m,i) - theta * col(m,i) */       

columnop(m,i,j, theta) := transpose(rowop(transpose(m),i, j, theta));

hipow_gzero(e,x) := block([n : hipow(e,x)], if n > 0 then n else 'inf);
good_pivot(e,x) := freeof(x,e) and e # 0;

ptriangularize(m,v) := block([p : ptriangularize_with_proviso(m,v)],
   if not(emptyp(p[2])) then print("Proviso: ",p[2]),
   p[1]); 
  
ptriangularize_with_proviso(m,v) := block([nr, nc, proviso : [], mp],
  require_unblockedmatrix(m, "first", "ptriangularize"),
  require_symbol(v, "second", "ptriangularize"),
  nr : length(m),
  nc : length(first(m)),
  for i : 1 thru min(nr, nc) do (
     mp : column_reduce(m,i,v),
     m : first(mp),
     proviso : append(proviso, second(mp))),
  proviso : setify(proviso),
  [expand(m), ratsimp(proviso)]);

column_reduce(m,i,x) := block([nc, nr, pos, q, proviso : []],
   /* require_matrix(m, "first", "column_reduce"), */
   nc : length(first(m)),
   nr : length(m),
   /* require_integer(i, "second", "column_reduce"), 
   if (i < 1) or (i > (length(first(m)))) then error("Bogus matrix column"), */
 
   while not(good_pivot(m[i,i],x)) and (i+1 <= nr) and
        listp(pos : locate_matrix_entry(m,i+1,i,nr,i, lambda([e],e # 0), 'bool)) do (
     pos : locate_matrix_entry(m,i,i,nr,i,lambda([e], good_pivot(e,x)), 'bool),
     if listp(pos) then (
       m : rowswap(m,i,pos[1]))
     else (
       m : expand(m),
       pos : locate_matrix_entry(m, i, i, nr, i, lambda([e], hipow(e,x)), 'max),
       m : rowswap(m,i,pos[1]),
       pos : locate_matrix_entry(m, i+1, i, nr, i, lambda([e], hipow_gzero(e,x)), 'min),
       if listp(pos) then (
         q : divide(m[i,i], m[pos[1],i],x),
         m : rowop(m,i,pos[1],first(q))))),
                
     pos : locate_matrix_entry(m, i, i, nr, i, lambda([e], good_pivot(e,x)), 'bool),

     if listp(pos) then (
        m : rowswap(m,i,first(pos)),
        for j : i+1 thru nr do (
          if not(constantp(ratnumer(m[i,i]))) then (
            proviso : cons(ratnumer(m[i,i]) # 0, proviso)),
           if not(constantp(ratdenom(m[i,i]))) then (
            proviso : cons(ratdenom(m[i,i]) # 0, proviso)),
          m : rowop(m,j,i,m[j,i]/m[i,i])))
     else (
       pos : locate_matrix_entry(m, i, i, nr, i, lambda([e], e # 0), 'bool),
       if listp(pos) then m : rowswap(m,i, first(pos))),
    [m, proviso]);

mat_norm(m, p) := block([listarith : true],
  require_matrix(m,"first","mat_norm"),
  if blockmatrixp(m) then m : mat_fullunblocker(m),
  if p = 1 then lmax(xreduce("+", map(lambda([s], map('cabs, s)), if args(m) = [] then [[]] else args(m))))
  else if p = 'inf then lmax(map(lambda([s], xreduce("+", map('cabs, s))), args(m)))	
  else if p = 'frobenius then sqrt(xreduce("+", lreduce('append, args(cabs(m)^2))))
  else error("Not able to compute the ",p," norm"));

mat_fullunblocker(m) := block(
  require_matrix(m, "first", "mat_unblocker"),
  if blockmatrixp(m) then (
     mat_fullunblocker(lreduce('addrow, args(map(lambda([x], lreduce('addcol, x)), m)))))
   else m);

mat_unblocker(m):=block(
  require_matrix(m, "first", "mat_unblocker"),
  if blockmatrixp(m) then (
     lreduce('addrow, args(map(lambda([x], lreduce('addcol, x)), m))))     
   else m);

mat_trace(m) := block([n, acc : 0],
  if not matrixp(m) then funmake('mat_trace, [m]) else (
    n : matrix_size(m),
    if first(n) # second(n) then error("The first argument to 'mat_trace' must be a square matrix"),
    n : first(n),
    for i from 1 thru n do (
      acc : acc + if matrixp(m[i,i]) then mat_trace(m[i,i]) else m[i,i]),
    acc));

kronecker_product(a,b) := (
  require_matrix(a,"first", "kronecker_product"),
  require_matrix(b,"second", "kronecker_product"),
  mat_unblocker(outermap('matrix_element_mult, a,b)));
  
diag_matrix([d]) := block([f, n, prederror : false],

  if every(lambda([s], matrixp(s) and not(blockmatrixp(s))), d) then (
    f : lambda([i,j], if i = j then inpart(d,i) 
	       else zeromatrix(first(matrix_size(inpart(d,i))), second(matrix_size(inpart(d,j))))))

  else if some(lambda([s], matrixp(s) or blockmatrixp(s)), d) then 
     error("All arguments to 'diag_matrix' must either be unblocked matrices or non-matrices")

  else f : lambda([i,j], if i = j then inpart(d,i) else 0),

  n : length(d),
  genmatrix(f, n, n));

hilbert_matrix(n) := block([row, mat : []],
  mode_declare(n,integer),
  require_posinteger(n,"first","hilbert_matrix"),
  row : makelist(1/i,i,1,n),
  for i : 1 thru n do (
    mat : cons(row, mat),
    row : endcons(1/(n + i), rest(row))),
  funmake('matrix, reverse(mat)));

hankel([q]) := block([col,row,m,n,partswitch : false],
  if length(q) > 2 or length(q) < 1 then error("The function 'hankel' requires one or two arguments"),
  col : inpart(q,1),
  row : if length(q) = 2 then inpart(q,2) else map(lambda([x],0), col),

  require_list(row,"first","hankel"),
  require_list(col,"second","hankel"),  
  m : length(row),
  n : length(col),
  genmatrix(lambda([i,j],if i+j-1 <= n then inpart(col,i+j-1) else inpart(row,i+j-n)),n,m));

toeplitz([q]) := block([col,row,m,n,partswitch : false],
  if length(q) > 2 or length(q) < 1 then error("The function 'toeplitz' requires one or two arguments"),
  col : inpart(q,1),
  row : if length(q) = 2 then inpart(q,2) else map('conjugate, col),

  require_list(row,"first","toeplitz"),
  require_list(col,"second","toeplitz"),  
  m : length(row),
  n : length(col),
  genmatrix(lambda([i,j], if i -j >= 0 then inpart(col, i-j+1) else inpart(row,j-i+1)),n,m));

polytocompanion(p,x) := block([n],
  if not polynomialp(p,[x], lambda([e], freeof(x,e))) then 
     error("First argument to 'polytocompanion' must be a polynomial"), 
  p : expand(p),
  n : hipow(p,x),
  p : multthru(p / coeff(p,x,n)),
  genmatrix(lambda([i,j], if i-j=1 then 1 else if j = n then -coeff(p,x,i-1) else 0),n));
      
moore_penrose_pseudoinverse(m) := block([mm, z : gensym(), scalarmatrixp : false],
  require_unblockedmatrix(m, "first", "moore_penrose_pseudoinverse"),
  mm : m . ctranspose(m),	
  limit(mm : ctranspose(m) . (mm - z * identfor(mm))^^-1,z,0));