Join Reorder Optimization - Abstract

Query Simplification: Graceful Degradation for Join-Order Optimization

This can refer to the related work, out of a more comprehensive,

FIG Query divided into the following several,

Graph Simplification algorithm,

Heuristic

Optimization of Large Join Queries: Combining Heuristic and Combinatorial Techniques

Main point of this article, and combine Combinatorial Heuristic

Combinatorial mean combination

Combinatorial optimization problem is to find an optimal state in the state space, cost the state is determined by the cost function

Combinatorial optimization algorithm is divided into two types,

Iterative algorithms, here mainly refers REPEAT, continuously random retry to find a more optimal solution

Annealing algorithm,

Heuristic algorithms,

First, Augmentation Heuristic

Initially only the first table,

Then one by one add up, plus function which depends chooseNext

Here are some indicators can be used as chooseNext, the paper said the results are the best 3 results

KBZ Heuristic

Algorithm is divided into three parts,

R, given a rooted tree, given optimal join ordering

T, given a join tree, through all the root, each rooted tree by R identify the optimal

G, given a join graph, may be cyclic, find a spanning tree (MST), call T

Local Improvement

Divide and rule, too many table number when the price is high exhaustive, but cut into small cluster, will be much simpler

The same would also not be the optimal solution, cluster can coincide

Finally, if the two technologies together?

II and SA is the two basic Combinatorial method,

SAA，SAK分别把augmentation和KBZ两种Heuristic方法用于SA，用于产生一个较优的initial state

IAI，IKI，用Heuristic的方法产生每一轮迭代的initial state

IAL，加入local improvement

AGI，KBI，先用Heuristic产生state，再用Iterative去优化

A New Heuristic for Optimizing Large Queries

查询优化的目的是避免worst plans，而不是找到best plan，在这样的假设下，启发式算法可能会达到比较好的效果

当前基于combinatorial优化技术(比如iterative或退火)的cost-based searching，已经取得了一定的效果，但是当前的方法并没有利用queries中inherent的semantic information

所以基本的思路就是，在当前cost-based searching的基础上利用semantic information，从而提出Goo算法，Greedy Operator Ordering

这是一种，Greedy的bottom up算法

Node关键属性是Size，Edge关键属性是Selectivity

Goo的目的是逐渐合并各个node，

合并的标准是，每次都是找产生中间结果最小的edge进行合并

很明显，Goo产生的肯定不是最优解

一般的思路都是，基于启发式的结果，进行进一步的调整和优化，找到更优解，

比如，增加一组rules，bottom up的试图apply这些rules得到更好的结果

Polynomial Heuristics for Query Optimization

One line of work adapts randomized techniques and combinatorial heuristics to address this problem.
These techniques consider the space of plans as points in a high-dimensional space, that can be “traversed” via transformations (e.g., join commutativity and
associativity).
Reference [13] surveys different such strategies, including iterative improvement, simulated annealing, and genetic algorithms.
These techniques can be seen as heuristic variations of transformation-based exhaustive enumeration algorithms.
Another line of work implements heuristic variations of dynamic programming. These approaches include reference [14] (which performs dynamic programming for a
subset of tables, picks the best k-table join, replaces it with a new “virtual” table, and repeats the procedure until all tables are part of the final plan),
reference [15] (which simplifies an initial join graph by disallowing non-promising join edges and then exhaustively searches the resulting, simpler problem using [8]), and references [16], [17] (which greedily build join trees one table at a time).

本文首先给出一个分类，比较新颖，

启发式是优化的基本技术，分为对于Transformation-based技术的启发式优化，和动态规划的启发式优化

其中Heuristic DP算法都是基于graph的，可以采用iterative的方式，根据cost等信息降低搜索空间等，或者用Greedy算法

但是文中说除了greedy的方案，其他的性能都太差

所以文中给出一个通用的Greedy算法框架，ERM