forked from aimacode/aima-julia
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathplanning.jl
1089 lines (990 loc) · 49 KB
/
planning.jl
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
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
export AbstractPDDL,
PDDL, goal_test, execute_action,
AbstractPlanningAction, PlanningAction,
substitute, check_precondition,
air_cargo_pddl, air_cargo_pddl_goal_test,
spare_tire_pddl, spare_tire_pddl_goal_test,
three_block_tower_pddl, three_block_tower_pddl_goal_test,
have_cake_and_eat_cake_too_pddl, have_cake_and_eat_cake_too_pddl_goal_test,
PlanningLevel,
find_mutex_links, build_level_links, build_level_links_permute_arguments, perform_actions,
planning_combinations,
PlanningGraph, expand_graph, non_mutex_goals,
GraphPlanProblem, check_level_off, extract_solution,
graphplan,
doubles_tennis_pddl, doubles_tennis_pddl_goal_test,
PlanningHighLevelAction, check_resource, check_job_order,
HighLevelPDDL, refinements, hierarchical_search,
job_shop_scheduling_pddl, job_shop_scheduling_pddl_goal_test;
abstract type AbstractPDDL end;
abstract type AbstractPlanningAction end;
#=
PlanningAction is an action schema defined by the action's name, preconditions, and effects.
Preconditions and effects consists of either positive and negated literals.
=#
struct PlanningAction <: AbstractPlanningAction
name::String
arguments::Tuple
precondition_positive::Array{Expression, 1}
precondition_negated::Array{Expression, 1}
effect_add_list::Array{Expression, 1}
effect_delete_list::Array{Expression, 1}
function PlanningAction(action::Expression, precondition::Tuple{Vararg{Array, 2}}, effect::Tuple{Vararg{Array, 2}})
return new(action.operator, action.arguments, precondition[1], precondition[2], effect[1], effect[2]);
end
end
function substitute(action::T, e::Expression, arguments::Tuple{Vararg{Expression}}) where {T <: AbstractPlanningAction}
local new_arguments::AbstractVector = collect(e.arguments);
for (index_1, argument) in enumerate(e.arguments)
for index_2 in 1:length(action.arguments)
if (action.arguments[index_2] == argument)
new_arguments[index_1] = arguments[index_2];
end
end
end
return Expression(e.operator, Tuple((new_arguments...,)));
end
function check_precondition(action::T1, kb::T2, arguments::Tuple) where {T1 <: AbstractPlanningAction, T2 <: AbstractKnowledgeBase}
# Check for positive clauses.
for clause in action.precondition_positive
if (!(substitute(action, clause, arguments) in kb.clauses))
return false;
end
end
# Check for negated clauses.
for clause in action.precondition_negated
if (substitute(action, clause, arguments) in kb.clauses)
return false;
end
end
return true;
end
"""
execute_action(action, kb, arguments)
Execute the given action 'action' on the knowledge base of the planning problem.
"""
function execute_action(action::T1, kb::T2, arguments::Tuple) where {T1 <: AbstractPlanningAction, T2 <: AbstractKnowledgeBase}
if (!(check_precondition(action, kb, arguments)))
error(@sprintf("execute_action(): Action \"%s\" preconditions are not satisfied!", action.name));
end
# Retract negated literals to knowledge base 'kb'.
for clause in action.effect_delete_list
retract(kb, substitute(action, clause, arguments));
end
# Add positive literals to knowledge base 'kb'.
for clause in action.effect_add_list
tell(kb, substitute(action, clause, arguments));
end
nothing;
end
#=
The Planning Domain Definition Language (PDDL) is used to define a search problem.
The states (starting from the initial state) are represented as the conjunction of
the statements in 'kb' (a FirstOrderLogicKnowledgeBase). The actions are described
by 'actions' (an array of action schemas). The 'goal_test' is a function that checks
if the current state of the problem is at the goal state.
=#
struct PDDL <: AbstractPDDL
kb::FirstOrderLogicKnowledgeBase
actions::Array{PlanningAction, 1}
goal_test::Function
function PDDL(initial_state::Array{Expression, 1}, actions::Array{PlanningAction, 1}, goal_test::Function)
return new(FirstOrderLogicKnowledgeBase(initial_state), actions, goal_test);
end
end
function goal_test(plan::T) where {T <: AbstractPDDL}
return plan.goal_test(plan.kb);
end
"""
execute_action(plan, action)
Execute the first relevant PlanningAction associated with action 'action'
on the given planning problem's knowledgebase.
"""
function execute_action(plan::T, action::Expression) where {T <: AbstractPDDL}
local action_name::String = action.operator;
local arguments::Tuple = action.arguments;
local relevant_actions::AbstractVector = collect(a for a in plan.actions if (a.name == action_name));
if (length(relevant_actions) == 0)
error(@sprintf("execute_action(): Action \"%s\" not found!", action_name));
else
local first_relevant_action::PlanningAction = relevant_actions[1];
if (!check_precondition(first_relevant_action, plan.kb, arguments))
error(@sprintf("execute_action(): Action \"%s\" preconditions are not satisfied!", repr(action)));
else
execute_action(first_relevant_action, plan.kb, arguments);
end
end
nothing;
end
function air_cargo_pddl_goal_test(kb::FirstOrderLogicKnowledgeBase)
return all((function(ans)
if (typeof(ans) <: Bool)
return ans;
else
if (length(ans) == 0) # length of Tuple
return false;
else
return true;
end
end
end),
collect(ask(kb, q) for q in (expr("At(C1, JFK)"), expr("At(C2, SFO)"))));
end
"""
air_cargo_pddl()
Return a PDDL representing the air cargo transportation planning problem (Fig. 10.1).
"""
function air_cargo_pddl()
local initial::Array{Expression, 1} = map(expr, ["At(C1, SFO)",
"At(C2, JFK)",
"At(P1, SFO)",
"At(P2, JFK)",
"Cargo(C1)",
"Cargo(C2)",
"Plane(P1)",
"Plane(P2)",
"Airport(JFK)",
"Airport(SFO)"]);
# Load Action Schema
local precondition_positive::Array{Expression, 1} = map(expr, ["At(c, a)",
"At(p, a)",
"Cargo(c)",
"Plane(p)",
"Airport(a)"]);
local precondition_negated::Array{Expression, 1} = [];
local effect_add_list::Array{Expression, 1} = [expr("In(c, p)")];
local effect_delete_list::Array{Expression, 1} = [expr("At(c, a)")];
local load::PlanningAction = PlanningAction(expr("Load(c, p, a)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
# Unload Action Schema
precondition_positive = map(expr, ["In(c, p)", "At(p, a)", "Cargo(c)", "Plane(p)", "Airport(a)"]);
precondition_negated = [];
effect_add_list = [expr("At(c, a)")];
effect_delete_list = [expr("In(c, p)")];
local unload::PlanningAction = PlanningAction(expr("Unload(c, p, a)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
# Fly Action Schema
precondition_positive = map(expr, ["At(p, f)", "Plane(p)", "Airport(f)", "Airport(to)"]);
precondition_negated = [];
effect_add_list = [expr("At(p, to)")];
effect_delete_list = [expr("At(p, f)")];
local fly::PlanningAction = PlanningAction(expr("Fly(p, f, to)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
return PDDL(initial, [load, unload, fly], air_cargo_pddl_goal_test);
end
function spare_tire_pddl_goal_test(kb::FirstOrderLogicKnowledgeBase)
return all((function(ans)
if (typeof(ans) <: Bool)
return ans;
else
if (length(ans) == 0) # length of Tuple
return false;
else
return true;
end
end
end),
collect(ask(kb, q) for q in (expr("At(Spare, Axle)"),)));
end
"""
spare_tire_pddl()
Return a PDDL representing the spare tire planning problem (Fig. 10.2).
"""
function spare_tire_pddl()
local initial::Array{Expression, 1} = map(expr, ["Tire(Flat)",
"Tire(Spare)",
"At(Flat, Axle)",
"At(Spare, Trunk)"]);
# Remove Action Schema
local precondition_positive::Array{Expression, 1} = [expr("At(obj, loc)")];
local precondition_negated::Array{Expression, 1} = [];
local effect_add_list::Array{Expression, 1} = [expr("At(obj, Ground)")];
local effect_delete_list::Array{Expression, 1} = [expr("At(obj, loc)")];
local remove::PlanningAction = PlanningAction(expr("Remove(obj, loc)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
# PutOn Action Schema
precondition_positive = map(expr, ["Tire(t)", "At(t, Ground)"]);
precondition_negated = [expr("At(Flat, Axle)")];
effect_add_list = [expr("At(t, Axle)")];
effect_delete_list = [expr("At(t, Ground)")];
local put_on::PlanningAction = PlanningAction(expr("PutOn(t, Axle)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
# LeaveOvernight Action Schema
precondition_positive = [];
precondition_negated = [];
effect_add_list = [];
effect_delete_list = map(expr, ["At(Spare, Ground)", "At(Spare, Axle)", "At(Spare, Trunk)",
"At(Flat, Ground)", "At(Flat, Axle)", "At(Flat, Trunk)"]);
local leave_overnight::PlanningAction = PlanningAction(expr("LeaveOvernight"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
return PDDL(initial, [remove, put_on, leave_overnight], spare_tire_pddl_goal_test);
end
function three_block_tower_pddl_goal_test(kb::FirstOrderLogicKnowledgeBase)
return all((function(ans)
if (typeof(ans) <: Bool)
return ans;
else
if (length(ans) == 0) #length of Tuple
return false;
else
return true;
end
end
end),
collect(ask(kb, q) for q in (expr("On(A, B)"), expr("On(B, C)"))));
end
"""
three_block_tower_pddl()
Return a PDDL representing the building of a three-block tower planning problem (Fig. 10.3).
"""
function three_block_tower_pddl()
local initial::Array{Expression, 1} = map(expr, ["On(A, Table)",
"On(B, Table)",
"On(C, A)",
"Block(A)",
"Block(B)",
"Block(C)",
"Clear(B)",
"Clear(C)"]);
# Move Action Schema
local precondition_positive::Array{Expression, 1} = map(expr, ["On(b, x)", "Clear(b)", "Clear(y)", "Block(b)", "Block(y)"]);
local precondition_negated::Array{Expression, 1} = [];
local effect_add_list::Array{Expression, 1} = [expr("On(b, y)"), expr("Clear(x)")];
local effect_delete_list::Array{Expression, 1} = [expr("On(b, x)"), expr("Clear(y)")];
local move::PlanningAction = PlanningAction(expr("Move(b, x, y)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
# MoveToTable Action Schema
precondition_positive = map(expr, ["On(b, x)", "Clear(b)", "Block(b)"]);
precondition_negated = [];
effect_add_list = [expr("On(b, Table)"), expr("Clear(x)")];
effect_delete_list = [expr("On(b, x)")];
local move_to_table::PlanningAction = PlanningAction(expr("MoveToTable(b, x)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
return PDDL(initial, [move, move_to_table], three_block_tower_pddl_goal_test);
end
function have_cake_and_eat_cake_too_pddl_goal_test(kb::FirstOrderLogicKnowledgeBase)
return all((function(ans)
if (typeof(ans) <: Bool)
return ans;
else
if (length(ans) == 0) # length of Tuple
return false;
else
return true;
end
end
end),
collect(ask(kb, q) for q in (expr("Have(Cake)"), expr("Eaten(Cake)"))));
end
"""
have_cake_and_eat_cake_too_pddl()
Return a PDDL representing the 'have cake and eat cake too' planning problem (Fig. 10.7).
"""
function have_cake_and_eat_cake_too_pddl()
local initial::Array{Expression, 1} = [expr("Have(Cake)")];
# Eat Cake Action Schema
local precondition_positive::Array{Expression, 1} = [expr("Have(Cake)")];
local precondition_negated::Array{Expression, 1} = [];
local effect_add_list::Array{Expression, 1} = [expr("Eaten(Cake)")];
local effect_delete_list::Array{Expression, 1} = [expr("Have(Cake)")];
local eat_cake::PlanningAction = PlanningAction(expr("Eat(Cake)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
# Bake Cake Action Schema
precondition_positive = [];
precondition_negated = [expr("Have(Cake)")];
effect_add_list = [expr("Have(Cake)")];
effect_delete_list = [];
local bake_cake::PlanningAction = PlanningAction(expr("Bake(Cake)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
return PDDL(initial, [eat_cake, bake_cake], have_cake_and_eat_cake_too_pddl_goal_test);
end
mutable struct PlanningLevel
positive_kb::FirstOrderLogicKnowledgeBase
current_state_positive::Array{Expression, 1} #current state of the planning problem
current_state_negated::Array{Expression, 1} #current state of the planning problem
current_action_links_positive::Dict #current actions to current state link
current_action_links_negated::Dict #current actions to current state link
current_state_links_positive::Dict #current state to action link
current_state_links_negated::Dict #current state to action link
next_action_links::Dict #current action to next state link
next_state_links_positive::Dict #next state to current action link
next_state_links_negated::Dict #next state to current action link
mutex_links::Array{Set, 1} #each mutex relation is a Set of 2 actions/literals
function PlanningLevel(p_kb::FirstOrderLogicKnowledgeBase, n_kb::FirstOrderLogicKnowledgeBase)
return new(p_kb, p_kb.clauses, n_kb.clauses, Dict(), Dict(), Dict(), Dict(), Dict(), Dict(), Dict(), []);
end
end
function find_mutex_links(level::PlanningLevel)
# Inconsistent effects condition between 2 action schemas at a given level
for positive_effect in level.next_state_links_positive
negated_effect = positive_effect;
if (haskey(level.next_state_links_negated, negated_effect))
for a in level.next_state_links_positive[positive_effect]
for b in level.next_state_links_negated[negated_effect]
if (!(Set([a, b]) in level.mutex_links))
push!(level.mutex_links, Set([a, b]));
end
end
end
end
end
# Inference condition between 2 action schemas at a given level
for positive_precondition in level.current_state_links_positive
negated_effect = positive_precondition;
if (haskey(level.next_state_links_negated, negated_effect))
for a in level.current_state_links_positive[positive_precondition]
for b in level.next_state_links_negated[negated_effect]
if (!(Set([a, b]) in level.mutex_links))
push!(level.mutex_links, Set([a, b]));
end
end
end
end
end
for negated_precondition in level.current_state_links_negated
positive_effect = negated_precondition;
if (haskey(level.next_state_links_positive, positive_effect))
for a in level.next_state_links_positive[positive_effect]
for b in level.current_state_links_negated[negated_precondition]
if (!(Set([a, b]) in level.mutex_links))
push!(level.mutex_links, Set([a, b]));
end
end
end
end
end
# Competing needs condition between 2 action schemas
for positive_precondition in level. current_state_links_positive
negated_precondition = positive_precondition;
if (haskey(level.current_state_links_negated, negated_precondition))
for a in level.current_state_links_positive[positive_precondition]
for b in level.current_state_links_negated[negated_precondition]
if (!(Set([a, b]) in level.mutex_links))
push!(level.mutex_links, Set([a, b]));
end
end
end
end
end
# Inconsistent support condition
local state_mutex_links::AbstractVector = [];
for pair in level.mutex_links
collected_pair::AbstractVector = collect(pair);
next_state_1 = level.next_action_links[collected_pair[1]];
if (length(sorted_pair) == 2)
next_state_2 = level.next_action_links[collected_pair[2]];
else
next_state_2 = level.next_action_links[collected_pair[1]];
end
if ((length(next_state_1) == 1) && (length(next_state_2) == 1))
push!(state_mutex_links, Set([next_state_1[1], next_state_2[1]]));
end
end
level.mutex_links = vcat(level.mutex_links, state_mutex_links);
nothing;
end
function build_level_links_permute_arguments(depth::Int64, objects::AbstractVector, current_permutation::Tuple, permutations_array::AbstractVector)
if (depth == 0)
push!(permutations_array, current_permutation);
elseif (depth < 0)
error("build_level_links_permute_arguments(): Found negative depth!");
else
for (i, item) in enumerate(objects)
build_level_links_permute_arguments((depth - 1),
Tuple((objects[1:(i - 1)]..., objects[(i + 1):end]...,)),
Tuple((current_permutation..., item)),
permutations_array)
end
end
end
function build_level_links_permute_arguments(depth::Int64, objects::Tuple, current_permutation::Tuple, permutations_array::AbstractVector)
if (depth == 0)
push!(permutations_array, current_permutation);
elseif (depth < 0)
error("build_level_links_permute_arguments(): Found negative depth!");
else
for (i, item) in enumerate(objects)
build_level_links_permute_arguments((depth - 1),
Tuple((objects[1:(i - 1)]..., objects[(i + 1):end]...,)),
Tuple((current_permutation..., item)),
permutations_array)
end
end
end
function build_level_links(level::PlanningLevel, actions::AbstractVector, objects::Set)
# Create persistence actions for positive states
for clause in level.current_state_positive
level.current_action_links_positive[Expression("Persistence", clause)] = [clause];
level.next_action_links[Expression("Persistence", clause)] = [clause];
level.current_state_links_positive[clause] = [Expression("Persistence", clause)];
level.next_state_links_positive[clause] = [Expression("Persistence", clause)];
end
# Create persistence actions for negated states
for clause in level.current_state_negated
not_expression = Expression("not"*clause.operator, clause.arguments);
level.current_action_links_negated[Expression("Persistence", not_expression)] = [clause];
level.next_action_links[Expression("Persistence", not_expression)] = [clause];
level.current_state_links_negated[clause] = [Expression("Persistence", not_expression)];
level.next_state_links_negated[clause] = [Expression("Persistence", not_expression)];
end
# Recursively collect num_arg depth, collecting a Tuple of Tuples
for action in actions
local num_arguments::Int64 = length(action.arguments);
local possible_arguments::AbstractVector = [];
build_level_links_permute_arguments(num_arguments, collect(objects), (), possible_arguments);
for argument in possible_arguments
if (check_precondition(action, level.positive_kb, argument))
for (number, symbol) in enumerate(action.arguments)
if (!all(islowercase, symbol.operator))
argument = Tuple((argument[1:(number - 1)]..., symbol, argument[(number + 1):end]...,));
end
end
local new_action::Expression = substitute(action, Expression(action.name, action.arguments), argument);
level.current_action_links_positive[new_action] = [];
level.current_action_links_negated[new_action] = [];
local new_clause::Expression;
for clause in action.precondition_positive
new_clause = substitute(action, clause, argument);
push!(level.current_action_links_positive[new_action], new_clause);
if (haskey(level.current_state_links_positive, new_clause))
push!(level.current_state_links_positive[new_clause], new_action);
else
level.current_state_links_positive[new_clause] = [new_action];
end
end
for clause in action.precondition_negated
new_clause = substitute(action, clause, argument);
push!(level.current_action_links_negated[new_action], new_clause);
if (haskey(level.current_state_links_negated, new_clause))
push!(level.current_state_links_negated[new_clause], new_action);
else
level.current_state_links_negated[new_clause] = [new_action];
end
end
level.next_action_links[new_action] = [];
for clause in action.effect_add_list
new_clause = substitute(action, clause, argument);
push!(level.next_action_links[new_action], new_clause);
if (haskey(level.next_state_links_positive, new_clause))
push!(level.next_state_links_positive[new_clause], new_action);
else
level.next_state_links_positive[new_clause] = [new_action];
end
end
for clause in action.effect_delete_list
new_clause = substitute(action, clause, argument);
push!(level.next_action_links[new_action], new_clause);
if (haskey(level.next_state_links_negated, new_clause))
push!(level.next_state_links_negated[new_clause], new_action);
else
level.next_state_links_negated[new_clause] = [new_action];
end
end
end
end
end
nothing;
end
function perform_actions(level::PlanningLevel)
local new_kb_positive::FirstOrderLogicKnowledgeBase = FirstOrderLogicKnowledgeBase(Array{Expression, 1}(collect(Set(collect(keys(level.next_state_links_positive))))));
local new_kb_negated::FirstOrderLogicKnowledgeBase = FirstOrderLogicKnowledgeBase(Array{Expression, 1}(collect(Set(collect(keys(level.next_state_links_negated))))));
return PlanningLevel(new_kb_positive, new_kb_negated);
end
function planning_combinations(items::AbstractVector) #ordered permutations of length 2
local combinations::AbstractVector = [];
for (i, a) in enumerate(items)
for b in items[(i + 1):end]
push!(combinations, (a, b));
end
end
return combinations;
end
#=
PlanningGraph is an implementation of a planning graph data structure.
The planning graph is organized into levels based on the given PDDL
and negated knowledge base.
=#
struct PlanningGraph
pddl::AbstractPDDL
levels::Array{PlanningLevel, 1}
objects::Set{Expression}
function PlanningGraph(pddl::T, n_kb::FirstOrderLogicKnowledgeBase) where {T <: AbstractPDDL}
return new(pddl, [PlanningLevel(pddl.kb, n_kb)], Set(collect(arg for clause in vcat(pddl.kb.clauses, n_kb.clauses) for arg in clause.arguments)));
end
end
function expand_graph(pg::PlanningGraph)
local last_level = pg.levels[length(pg.levels)];
build_level_links(last_level, pg.pddl.actions, pg.objects);
find_mutex_links(last_level);
push!(pg.levels, perform_actions(last_level));
# for (i, level) in enumerate(pg.levels)
# println("Expanded graph level $i: ", level)
# end
nothing;
end
function non_mutex_goals(pg::PlanningGraph, goals::AbstractVector, index::Int64)
local goal_combinations::AbstractVector = planning_combinations(goals);
for goal in goal_combinations
if (index < 0)
if (Set(collect(goal)) in reverse(pg.levels)[abs(index)].mutex_links)
return false;
end
else
error("non_mutex_goals(): Expected negative index, got ", index, "!");
end
end
return true;
end
#=
GraphPlanProblem is a data structure that stores the 'graph', 'nogoods',
and 'solution' variables of a given Graphplan planning problem.
=#
struct GraphPlanProblem
graph::PlanningGraph
nogoods::AbstractVector
solution::AbstractVector
function GraphPlanProblem(pddl::AbstractPDDL, n_kb::FirstOrderLogicKnowledgeBase)
return new(PlanningGraph(pddl, n_kb), [], []);
end
end
function check_level_off(gpp::GraphPlanProblem)
local first_check::Bool = (Set(reverse(gpp.graph.levels)[1].current_state_positive) == Set(reverse(gpp.graph.levels)[2].current_state_positive));
local second_check::Bool = (Set(reverse(gpp.graph.levels)[1].current_state_negated) == Set(reverse(gpp.graph.levels)[2].current_state_negated));
return (first_check && second_check);
end
function extract_solution(gpp::GraphPlanProblem, goals_positive::AbstractVector, goals_negated::AbstractVector, index::Int64)
local level::PlanningLevel;
if (index < 0)
level = reverse(gpp.graph.levels)[abs(index)];
else
error("extract_solution(): Expected negative index, got ", index, "!");
end
if (!non_mutex_goals(gpp.graph, vcat(goals_positive, goals_negated), index))
push!(gpp.nogoods, (level, goals_positive, goals_negated));
return nothing;
end
level = reverse(gpp.graph.levels)[abs(index - 1)];
local actions::AbstractVector = [];
for goal in goals_positive
push!(actions, level.next_state_links_positive[goal]);
end
for goal in goals_negated
push!(actions, level.next_state_links_negated[goal]);
end
# Create all possible combinations of actions by using finding the cartesian product.
local action_combinations::AbstractVector = iterable_cartesian_product(actions);
# Remove action combinations that contain mutexes.
local non_mutex_actions::AbstractVector = [];
for action_list in action_combinations
action_pairs = planning_combinations(collect(Set(action_list)));
push!(non_mutex_actions, collect(Set(action_list)));
for pair in action_pairs
if (Set(collect(pair)) in level.mutex_links)
pop!(non_mutex_actions);
break;
end
end
end
for action_list in non_mutex_actions
if (!([action_list, index] in gpp.solution))
push!(gpp.solution, [action_list, index]);
local new_goals_positive::AbstractVector = [];
local new_goals_negated::AbstractVector = [];
for action in Set(action_list)
if (haskey(level.current_action_links_positive, action))
new_goals_positive = vcat(new_goals_positive, level.current_action_links_positive[action]);
end
if (haskey(level.current_action_links_negated, action))
new_goals_negated = vcat(new_goals_negated, level.current_action_links_negated[action]);
end
end
if ((abs(index) + 1) == length(gpp.graph.levels))
return nothing;
elseif ((level, new_goals_positive, new_goals_negated) in gpp.nogoods)
return nothing;
else
extract_solution(gpp, new_goals_positive, new_goals_negated, (index - 1))
end
end
end
local solution::AbstractVector = [];
for item in gpp.solution
if (item[2] == -1)
push!(solution, [item[1]]);
else
push!(reverse(solution)[1], item[1]);
end
end
for (i, item) in enumerate(solution)
solution[i] = reverse(item);
end
return solution;
end
function goal_test(gpp::GraphPlanProblem, goals::AbstractVector) #goal_test() for graphplan()
local positive_kb::FirstOrderLogicKnowledgeBase = reverse(gpp.graph.levels)[1].positive_kb;
return all((function(ans)
if (typeof(ans) <: Bool)
return ans;
else
if (length(ans) == 0) # length of Tuple
return false;
else
return true;
end
end
end),
collect(ask(positive_kb, goal) for goal in goals));
end
"""
graphplan(gpp::GraphPlanProblem, goals::Tuple)
Apply the Graphplan algorithm (Fig. 10.9) to the given planning problem 'gpp' and goal state 'goals'.
Return the solution or 'nothing' on failure.
"""
function graphplan(gpp::GraphPlanProblem, goals::Tuple)
local goals_positive::AbstractVector = goals[1];
local goals_negated::AbstractVector = goals[2];
while (true)
if (goal_test(gpp, goals_positive) && non_mutex_goals(gpp.graph, vcat(goals_positive, goals_negated), -1))
solution = extract_solution(gpp, goals_positive, goals_negated, -1);
# println("Solution: ", solution)
if (!(typeof(solution) <: Nothing))
return solution;
end
end
expand_graph(gpp.graph);
if ((length(gpp.graph.levels) > 1) && check_level_off(gpp))
return nothing;
end
end
return nothing;
end
function doubles_tennis_pddl_goal_test(kb::FirstOrderLogicKnowledgeBase)
return all((function(ans)
if (typeof(ans) <: Bool)
return ans;
else
if (length(ans) == 0) #length of Tuple
return false;
else
return true;
end
end
end),
collect(ask(kb, q) for q in map(expr, ["Returned(Ball)", "At(a, RightNet)", "At(a, LeftNet)"])));
end
"""
doubles_tennis_pddl()
Return a PDDL representing the doubles tennis planning problem (Fig. 11.10).
The PlanningActions 'hit' and 'go' both take 3 arguments because their
preconditions will not be satisfied without the 'loc' variable.
This occurs when substituting propositional logic variables within
the check_precondition() function.
"""
function doubles_tennis_pddl()
local initial::Array{Expression, 1} = map(expr, ["At(A, LeftBaseLine)",
"At(B, RightNet)",
"Approaching(Ball, RightBaseLine)",
"Partner(A, B)",
"Partner(B, A)"]);
# Hit Action Schema
local precondition_positive::Array{Expression, 1} = map(expr, ["Approaching(Ball, loc)", "At(actor, loc)"]);
local precondition_negated::Array{Expression, 1} = [];
local effect_add_list::Array{Expression, 1} = [expr("Returned(Ball)")];
local effect_delete_list::Array{Expression, 1} = [];
local hit::PlanningAction = PlanningAction(expr("Hit(actor, loc, Ball)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
# Go Action Schema
precondition_positive = [expr("At(actor, loc)")];
precondition_negated = [];
effect_add_list = [expr("At(actor, to)")];
effect_delete_list = [expr("At(actor, loc)")];
local go::PlanningAction = PlanningAction(expr("Go(actor, loc, to)"),
(precondition_positive, precondition_negated),
(effect_add_list, effect_delete_list));
return PDDL(initial, [hit, go], doubles_tennis_pddl_goal_test);
end
#=
PlanningHighLevelAction (HLA) is an action schema with resource constraints and refinements.
=#
mutable struct PlanningHighLevelAction <: AbstractPlanningAction
name::String
arguments::Tuple
precondition_positive::Array{Expression, 1}
precondition_negated::Array{Expression, 1}
effect_add_list::Array{Expression, 1}
effect_delete_list::Array{Expression, 1}
duration::Int64
consumes::Dict
uses::Dict
completed::Bool
priority::Int64 #undefined field used to schedule multiple HLAs
job_group::Int64 #undefined field used to schedule multiple HLAs
function PlanningHighLevelAction(action::Expression,
precondition::Tuple{Vararg{Array{Expression, 1}, 2}},
effect::Tuple{Vararg{Array{Expression, 1}, 2}};
duration::Int64=0,
consumes::Dict=Dict(),
uses::Dict=Dict())
return new(action.operator, action.arguments, precondition[1], precondition[2], effect[1], effect[2], duration, consumes, uses, false);
end
end
"""
execute_action(hla, job_order, available_resources, kb, arguments)
Execute the given action 'hla' on the knowledge base of the planning problem.
This function checks for resources and job order execution required for the action to be executed.
"""
function execute_action(hla::PlanningHighLevelAction, job_order::Array{Array{PlanningHighLevelAction, 1}, 1}, available_resources::Dict, kb::AbstractKnowledgeBase, arguments::Tuple)
if (!check_consumes_resource(hla, available_resources))
error(@sprintf("execute_action(): High-level action \"%s\"'s consumable constraint was not satisfied!", hla.name));
elseif (!check_uses_resource(hla, available_resources))
error(@sprintf("execute_action(): High-level action \"%s\"'s reusable constraint was not satisfied!", hla.name));
elseif (!check_job_order(hla, job_order))
error(@sprintf("execute_action(): High-level action \"%s\" is not scheduled to execute! All scheduled unexecuted actions should run first!", hla.name));
end
# Execute 'hla' like a primitive action within this method to avoid confusion of
# calling execute_action() within execute_action().
if (!(check_precondition(hla, kb, arguments)))
error(@sprintf("execute_action(): High-level action \"%s\" preconditions are not satisfied!", hla.name));
end
# Retract negated literals to knowledge base 'kb'.
for clause in hla.effect_delete_list
retract(kb, substitute(hla, clause, arguments));
end
# Add positive literals to knowledge base 'kb'.
for clause in hla.effect_add_list
tell(kb, substitute(hla, clause, arguments));
end
for resource in keys(hla.consumes)
available_resources[resource] = available_resources[resource] - hla.consumes[resource];
end
# Update current high-level action task as completed.
hla.completed = true;
nothing;
end
function check_consumes_resource(hla::PlanningHighLevelAction, available_resources::Dict)
for resource in keys(hla.consumes)
if (!haskey(available_resources, resource))
return false;
elseif (available_resources[resource] < (hla.consumes)[resource])
return false;
end
end
return true;
end
function check_uses_resource(hla::PlanningHighLevelAction, available_resources::Dict)
for resource in keys(hla.uses)
if (!haskey(available_resources, resource))
return false;
elseif (available_resources[resource] < (hla.uses)[resource])
return false;
end
end
return true;
end
function check_job_order(hla::PlanningHighLevelAction, job_order::Array{Array{PlanningHighLevelAction, 1}, 1})
for jobs in job_order
if (hla in jobs)
for job in jobs
if (job == hla)
return true;
elseif (!job.completed)
return false;
end
end
end
end
return true;
end
#=
HighLevelPDDL is a high-level Planning Domain Definition Language (PDDL) and is used to
define a search problem. The HighLevelPDDL includes job scheduling and resource constraints
in addition to the states, action schemas, and the goal test found in the PDDL datatype.
=#
struct HighLevelPDDL <: AbstractPDDL
kb::FirstOrderLogicKnowledgeBase
actions::Array{PlanningHighLevelAction, 1}
goal_test::Function
jobs::Array{Array{PlanningHighLevelAction, 1}, 1}
resources::Dict
function HighLevelPDDL(initial_state::Array{Expression, 1}, actions::Array{PlanningHighLevelAction, 1}, goal_test::Function; jobs::Array{Array{PlanningHighLevelAction, 1}, 1}=Array{Array{PlanningHighLevelAction, 1}, 1}(), resources::Dict=Dict())
return new(FirstOrderLogicKnowledgeBase(initial_state), actions, goal_test, jobs, resources);
end
end
"""
execute_action(plan, action)
Execute the first relevant PlanningHighLevelAction associated with action 'action'
on the given planning problem's knowledge base.
"""
function execute_action(plan::HighLevelPDDL, action::PlanningHighLevelAction)
local arguments::Tuple = action.arguments;
local relevant_actions::AbstractVector = collect(a for a in plan.actions if (a.name == action.name));
if (length(relevant_actions) == 0)
error(@sprintf("execute_action(): High-level action \"%s\" not found!", action.name));
else
local first_relevant_action::PlanningHighLevelAction = relevant_actions[1];
execute_action(first_relevant_action, plan.jobs, plan.resources, plan.kb, arguments);
end
nothing;
end
"""
get_result(plan, action)
Return the planning problem after executing the given high-level action 'action'.
"""
function get_result(plan::HighLevelPDDL, action::PlanningHighLevelAction)
execute_action(plan, action);
return plan;
end
function get_result(plan::HighLevelPDDL, action::Nothing)
return plan;
end
"""
refinements(hla, state, dict)
Return a action sequence such that each refinement's preconditions of 'hla' are satisfied by
the given state of the planning problem 'state'.
"""
function refinements(hla::PlanningHighLevelAction, state::HighLevelPDDL, dict::Dict)
local solution::Array{PlanningHighLevelAction, 1} = Array{PlanningHighLevelAction, 1}();
local indices::AbstractVector = collect(i for (i, x) in enumerate(dict["HLA"]) if (expr(x).operator == hla.name));
for i in indices
local action::PlanningHighLevelAction = PlanningHighLevelAction(expr(dict["steps"][i][1]),
(collect(expr(x) for x in dict["precondition_positive"][i]), collect(expr(x) for x in dict["precondition_negated"][i])),
(collect(expr(x) for x in dict["effect_add_list"][i]), collect(expr(x) for x in dict["effect_delete_list"][i])));
if (check_precondition(hla, state.kb, action.arguments))
push!(solution, action);
end
end
return solution;
end
"""
hierarchical_search(problem, hierarchy)
Use the breadth-first implementation of the Hierarchical Search algorithm (Fig. 11.5) to the
given high-level planning problem 'problem' and 'hierarchy'. Return the solution or 'nothing'
on failure.
"""
function hierarchical_search(problem::HighLevelPDDL, hierarchy)
local act::Node = Node(problem.actions[1]);
local frontier::FIFOQueue();
push!(frontier, act);
while (true)
if (length(frontier) == 0)
return nothing;
end
local plan = pop!(frontier);
hla = plan.state;
if (!(plan.parent === nothing))
prefix = plan.parent.action;
else
prefix = nothing;
end
local outcome::HighLevelPDDL = get_result(problem, prefix);
if (typeof(hla) <: Nothing)
if (goal_test(outcome))
return path(plan);
end
else
for sequence in refinements(hla, outcome, hierachy)
push!(frontier, Node(plan.state, parent=plan.parent, action=sequence));
end
end
end
nothing;
end
function job_shop_scheduling_pddl_goal_test(kb::FirstOrderLogicKnowledgeBase)