Algorithm Analysis Big O CS341 Dick Steflik Complexity

Algorithm Analysis (Big O) CS-341 Dick Steflik

Complexity • In examining algorithm efficiency we must understand the idea of complexity – Space complexity – Time Complexity

Space Complexity • when memory was expensive and machines didn’t have much, we focused on making programs as space efficient as possible and developed schemes to make memory appear larger than it really was (virtual memory and memory paging schemes) • Although not as important today, space complexity is still important in the field of embedded computing (hand held computer based equipment like cell phones, palm devices, etc)

Time Complexity • Is the algorithm “fast enough” for my needs • How much longer will the algorithm take if I increase the amount of data it must process • Given a set of algorithms that accomplish the same thing, which is the right one to choose

Cases to examine • Best case – if the algorithm is executed, the fewest number of instructions are executed • Average case – executing the algorithm produces path lengths that will on average be the same • Worst case – executing the algorithm produces path lengths that are always a maximum

Worst case analysis • Of the three cases the really only useful case (from the standpoint of program design) is that of the worst case. • Worst case helps answer the software lifecycle issue of: – If its good enough today, will it be good enough tomorrow

Frequency Count • examine a piece of code and predict the number of instructions to be executed • Ex for each instruction predict how mant times each will be encountered as the code runs Inst # 1 Code F. C. for (int i=0; i< n ; i++) n+1 2 { cout << i; n 3 p = p + i; n } ____ 3 n+1 totaling the counts produces the F. C. (frequency count)

Order of magnitude • In the previous example best_case=avg_case=worst_case because the example was based just on fixed iteration • taken by itself, F. C. is relatively meaningless but expressed as an order of magnitude we can use it as a estimator of algorithm performance as we increase the amount of data • to convert F. C. to order of magnitude: – discard constant terms – disregard coefficients – pick the most significant term • the order of magnitude of 3 n+1 becomes n • if F. C. is always calculated taking a worst case path through the algorithm the order of magnitude is called Big O (i. e. O(n))

Another example Inst # 1 2 Code F. C. for (int i=0; i< n ; i++) n+1 n(n+1) n 2+2 n+1 for int j=0 ; j < n; j++) 3 { cout << i; n*n n 2 4 p = p + i; n*n n 2 ____ 3 n+1 3 n 2+3 n+2 } discarding constant terms produces : clearing coefficients : n 2+n picking the most significant term: n 2 3 n 2+3 n Big O = O(n 2)

What is Big O • Big O is the rate at which performance of an algorithm degrades as a function of the amount of data it is asked to handle • For example: O(n) indicates that performance degrades at a linear rate; O(n 2) indicates the rate of degradation follows a quadratic path.

Common growth rates

Small collections of data n 2 log 2 n For small collections of data, an algorithm with a worse overall big O may actually perform better than an algorithm with a better big O 20

Combining Algorithms • Sometimes we can increase the overall performance of an algorithm (but not it’s big. O) by combining two algorithms. • Suppose we have a large number of data elements to be sorted and the algorithm we pick has O(log 2 n) but we notice that it spends a lot of time doing housekeeping for small sub-collections of data. • We can take advantage of the better performance of an O(n 2) algorithm on small collections of data by switching algorithms when we notice that the collection size is getting small • This will improve the overall performance (time that this specific case takes to run) but will not change the big. O

Summary • Order-of-magnitude analysis and Big O measure an algorithm’s time requirement as a function of of the problem size by using a growth rate function. This allows you to analyze the efficiency without regard to factors like the computer speed and skill of the person coding the solution. • When comparing the efficiency of algorithms, you examine their growth rate functions when the problems are large. Only significant differences in growth-rate functions are meaningful. • Worst case analysis considers the maximum amount of work an algorithm will require while average case analysis considers the expected amount of work. • Use order-of-magnitude analysis to help you select a particular implementation for an ADT. If your application frequently uses particular ADT operations, your implementation should be efficient for at least those operations.