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
|
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>62.3. Genetic Query Optimization (GEQO) in PostgreSQL</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="geqo-intro2.html" title="62.2. Genetic Algorithms" /><link rel="next" href="geqo-biblio.html" title="62.4. Further Reading" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">62.3. Genetic Query Optimization (<acronym class="acronym">GEQO</acronym>) in PostgreSQL</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="geqo-intro2.html" title="62.2. Genetic Algorithms">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="geqo.html" title="Chapter 62. Genetic Query Optimizer">Up</a></td><th width="60%" align="center">Chapter 62. Genetic Query Optimizer</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 15.7 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="geqo-biblio.html" title="62.4. Further Reading">Next</a></td></tr></table><hr /></div><div class="sect1" id="GEQO-PG-INTRO"><div class="titlepage"><div><div><h2 class="title" style="clear: both">62.3. Genetic Query Optimization (<acronym class="acronym">GEQO</acronym>) in PostgreSQL</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="geqo-pg-intro.html#id-1.10.13.5.6">62.3.1. Generating Possible Plans with <acronym class="acronym">GEQO</acronym></a></span></dt><dt><span class="sect2"><a href="geqo-pg-intro.html#GEQO-FUTURE">62.3.2. Future Implementation Tasks for
<span class="productname">PostgreSQL</span> <acronym class="acronym">GEQO</acronym></a></span></dt></dl></div><p>
The <acronym class="acronym">GEQO</acronym> module approaches the query
optimization problem as though it were the well-known traveling salesman
problem (<acronym class="acronym">TSP</acronym>).
Possible query plans are encoded as integer strings. Each string
represents the join order from one relation of the query to the next.
For example, the join tree
</p><pre class="literallayout">
/\
/\ 2
/\ 3
4 1
</pre><p>
is encoded by the integer string '4-1-3-2',
which means, first join relation '4' and '1', then '3', and
then '2', where 1, 2, 3, 4 are relation IDs within the
<span class="productname">PostgreSQL</span> optimizer.
</p><p>
Specific characteristics of the <acronym class="acronym">GEQO</acronym>
implementation in <span class="productname">PostgreSQL</span>
are:
</p><div class="itemizedlist"><ul class="itemizedlist compact" style="list-style-type: bullet; "><li class="listitem" style="list-style-type: disc"><p>
Usage of a <em class="firstterm">steady state</em> <acronym class="acronym">GA</acronym> (replacement of the least fit
individuals in a population, not whole-generational replacement)
allows fast convergence towards improved query plans. This is
essential for query handling with reasonable time;
</p></li><li class="listitem" style="list-style-type: disc"><p>
Usage of <em class="firstterm">edge recombination crossover</em>
which is especially suited to keep edge losses low for the
solution of the <acronym class="acronym">TSP</acronym> by means of a
<acronym class="acronym">GA</acronym>;
</p></li><li class="listitem" style="list-style-type: disc"><p>
Mutation as genetic operator is deprecated so that no repair
mechanisms are needed to generate legal <acronym class="acronym">TSP</acronym> tours.
</p></li></ul></div><p>
</p><p>
Parts of the <acronym class="acronym">GEQO</acronym> module are adapted from D. Whitley's
Genitor algorithm.
</p><p>
The <acronym class="acronym">GEQO</acronym> module allows
the <span class="productname">PostgreSQL</span> query optimizer to
support large join queries effectively through
non-exhaustive search.
</p><div class="sect2" id="id-1.10.13.5.6"><div class="titlepage"><div><div><h3 class="title">62.3.1. Generating Possible Plans with <acronym class="acronym">GEQO</acronym></h3></div></div></div><p>
The <acronym class="acronym">GEQO</acronym> planning process uses the standard planner
code to generate plans for scans of individual relations. Then join
plans are developed using the genetic approach. As shown above, each
candidate join plan is represented by a sequence in which to join
the base relations. In the initial stage, the <acronym class="acronym">GEQO</acronym>
code simply generates some possible join sequences at random. For each
join sequence considered, the standard planner code is invoked to
estimate the cost of performing the query using that join sequence.
(For each step of the join sequence, all three possible join strategies
are considered; and all the initially-determined relation scan plans
are available. The estimated cost is the cheapest of these
possibilities.) Join sequences with lower estimated cost are considered
<span class="quote">“<span class="quote">more fit</span>”</span> than those with higher cost. The genetic algorithm
discards the least fit candidates. Then new candidates are generated
by combining genes of more-fit candidates — that is, by using
randomly-chosen portions of known low-cost join sequences to create
new sequences for consideration. This process is repeated until a
preset number of join sequences have been considered; then the best
one found at any time during the search is used to generate the finished
plan.
</p><p>
This process is inherently nondeterministic, because of the randomized
choices made during both the initial population selection and subsequent
<span class="quote">“<span class="quote">mutation</span>”</span> of the best candidates. To avoid surprising changes
of the selected plan, each run of the GEQO algorithm restarts its
random number generator with the current <a class="xref" href="runtime-config-query.html#GUC-GEQO-SEED">geqo_seed</a>
parameter setting. As long as <code class="varname">geqo_seed</code> and the other
GEQO parameters are kept fixed, the same plan will be generated for a
given query (and other planner inputs such as statistics). To experiment
with different search paths, try changing <code class="varname">geqo_seed</code>.
</p></div><div class="sect2" id="GEQO-FUTURE"><div class="titlepage"><div><div><h3 class="title">62.3.2. Future Implementation Tasks for
<span class="productname">PostgreSQL</span> <acronym class="acronym">GEQO</acronym></h3></div></div></div><p>
Work is still needed to improve the genetic algorithm parameter
settings.
In file <code class="filename">src/backend/optimizer/geqo/geqo_main.c</code>,
routines
<code class="function">gimme_pool_size</code> and <code class="function">gimme_number_generations</code>,
we have to find a compromise for the parameter settings
to satisfy two competing demands:
</p><div class="itemizedlist"><ul class="itemizedlist compact" style="list-style-type: disc; "><li class="listitem"><p>
Optimality of the query plan
</p></li><li class="listitem"><p>
Computing time
</p></li></ul></div><p>
</p><p>
In the current implementation, the fitness of each candidate join
sequence is estimated by running the standard planner's join selection
and cost estimation code from scratch. To the extent that different
candidates use similar sub-sequences of joins, a great deal of work
will be repeated. This could be made significantly faster by retaining
cost estimates for sub-joins. The problem is to avoid expending
unreasonable amounts of memory on retaining that state.
</p><p>
At a more basic level, it is not clear that solving query optimization
with a GA algorithm designed for TSP is appropriate. In the TSP case,
the cost associated with any substring (partial tour) is independent
of the rest of the tour, but this is certainly not true for query
optimization. Thus it is questionable whether edge recombination
crossover is the most effective mutation procedure.
</p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="geqo-intro2.html" title="62.2. Genetic Algorithms">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="geqo.html" title="Chapter 62. Genetic Query Optimizer">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="geqo-biblio.html" title="62.4. Further Reading">Next</a></td></tr><tr><td width="40%" align="left" valign="top">62.2. Genetic Algorithms </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 15.7 Documentation">Home</a></td><td width="40%" align="right" valign="top"> 62.4. Further Reading</td></tr></table></div></body></html>
|
