Binary search tree

All ring chemistry tables are organized into what is called binary search trees. Queries are posed about the existence of certain structural features. Each of these questions is answerable with a yes or a no. Based on the answer one of two different followup questions is selected. Embedded within the table may be requests for subgoals, either those already in the PGI or PGA table or for special reactions which are needed only for these transforms and are not of general synthetic interest. 

The example below shows some of the synthetic routes generated by the Diels Alder transform for the indicated precursor. It is important to note that while some of the chemistry may look somewhat naive, it can be quite thought provoking. 

ACS Symposium Series Ameriean Chemieal Soeiety Washington, DC, 1977. 

In terms of newer chemistry, the sample sequence shown below was generated by the iodo-lactonization transform. 

---

Binary search tree Data structure used in a search. 

Descriptions of the simplest data structure entities and explanations of their nature follow in succeeding sections. Basic data structures are stack, queue, and other linear lists multiple-dimension arrays (recursive) lists and trees (including forests and binary trees). Pointer or link simply means computer data constituting a memory location. Level indicates position in a structure that is hierarchical. Link, level, and the elementary structures are almost intuitive concepts. They are fairly easily understood by reference to their names or to real-life situations to which they relate. Evolving computer practice has had two effects. First, the impact of the World Wide Web and Internet browsers has acquainted many computer users with two basic ideas link (pointer) and level. Second, computer specialists have increased their use of advanced data structures. These may be understandable from their names or descriptive properties. Some of these terms are tries, quad-trees (quadtrees, quaternary trees), leftist-trees, 2-3 trees, binary search trees, and heap. While they are less common data structures and unlikely to be part of a first course in the field, they enable algorithmic procedures in applications such as image transmission, geographic data, and library search. 

The many special cases of such recursively defined binary search trees (data structure) each correspond to a search algorithm. Movement through a binary search free is like going to a place in the file. That place has a key, which is to be examined and compared to the one sought. 

Two special cases of binary search trees (and, hence, search algorithms) are binary and Fibonaccian. Each of these algorithms is relatively easy to implement in terms of computations used to create the underlying search tree data structure. By contrast, interpolation search is a valuable method but is more complex in terms of computations. 

In a Fibonaccian search, the elements of the binary search tree are either Fibonacci numbers or derived from them the root is a Fibonacci number, as are all nodes reached by only left links. Right links lead to nodes whose values are the ancestor plus the difference between it and its left successor. That is, the difference between the ancestor and left successor is added to the ancestor to get the right successor value. Fibonaccian binary search trees have a total number of elements one less than a Fibonacci number. 

The formulation for this scenario entails 1411 constraints, 511 continuous and 120 binary variables. The reduction in continuous variables compared to scenario 1 is due to the absence of linearization variables, since no attempt was made to linearize the scenario 2 model as explained in Section 4.3. An average of 1100 nodes were explored in the branch and bound search tree during the three major iterations between the MILP master problem and the NLP subproblem. The problem was solved in 6.54 CPU seconds resulting in an optimal objective of 2052.31 kg, which corresponds to 13% reduction in freshwater requirement. The corresponding water recycle/reuse network is shown in Fig. 4.10. 

The corresponding mathematical formulation entails 5534 constraints, 1217 continuous and 280 binary variables. An average of 4000 nodes were explored in the branch and bound search tree. The solution required three major iterations and took 309.41 CPU seconds to obtain the optimal solution of 1285.50 kg. This corresponds to 45.53% reduction in freshwater demand. A water reuse/recycle network that corresponds to this solution is shown in Fig. 4.11. 

The overall model for this scenario involves 5614 constraints, 1132 continuous 280 binary variables. Three major iterations with an average of 1200 nodes in the branch and bound search tree were required in the solution. The objective value of 1560 kg, which corresponds to 33.89% reduction in freshwater requirement, was obtained in 60.24 CPU seconds. An equivalent of this scenario, without reusable water storage, i.e. scenario 2, resulted in 13% reduction in fresh water. Figure 4.12 shows the water recycle/reuse network corresponding to this solution. 

More generally, MILPs are solved with branch and bound algorithms, similar to the spatial branch and bound method of the previous section, that explore the search space. As seen in Fig. 3-61, binary variables are used to define the search tree, and a number of bounding properties can be noted from the structure of (3-110). 

In a parent tree data structure, each successor points to its ancestor. Hence, such a structure can be stored in memory as a sequential list of (node, parent-link) pairs, as illustrated by Fig. 3. The parent tree representation facilitates bottom-up operations, such as finding the (1) root, (2) depth in the tree, and (3) ancestors (i.e., all nodes in the chain from the selected one to the root). Another advantage is in savings in link overhead Only one link per node is required, compared to two per node in the conventional (downward-pointer binary tree) representation. The disadvantage of the parent representation is that it is inefficient for problems requiring either enumeration of all nodes in a tree or top-down exploration of tree sections. Further, it is valid only for nonordered trees. Trees where sibling order is not represented are less versatile data structures. For example, search trees cannot be represented as parent trees, since the search is guided by the order on keys stored in the data structure information fields. 

This is an infinite full binary tree. Roughly speaking, a single pushdown store cannot search such a potentially infinite tree. When we did so in Chapter VI, we added a counter whose length could be compared with the length of the store. In fact we cannot search such a tree with any number of pushdown stores unless they can find their "bottoms" - unless one can find out if the store is empty. The pushdown stores in this chapter do not have this ability. Hence leaftest cannot be done with any number of pushdown stores. However, two pushdown stores suffice if a special test "Is pushdown store i empty " is allowed, since in that case we can use two stores to simulate the action of the array in the array augmented scheme for leaftest. 

Figure 5.2 Binary tree for depth first search with backtracking...

The binary trees for (i) depth first search with backtracking and (ii) breadth first search are shown in Figures 5.2 and 5.3 respectively. The number within the nodes indicate the sequence of candidate subproblems for each type of search. 

Figure 5.3 Binary tree for breadth first search...

If even tighter bounds are required, then a search further down the binary tree of submatrices can be undertaken, starting from those matrices whose norm is greater than the largest eigenvalue so far, e. A 2"-ary tree may in fact be more efficient, because each matrix with norm Z > e need only be multiplied by those matrices with norm greater than e2/Z. 

It is contained in our real library and its MIC is 0.06. Following [273] we considered all compounds that have MIC < 0.06 as active. Our QSAR search is now a binary classification problem. We solved it using various methods of descriptor selection and classification methods. Growing classification trees, we obtain a 3-descriptor CT (Figure 7.24)... 

---