Data structure and algorithm analysis
From
(Difference between revisions)
Current revision as of 11:12, 21 September 2019
Justify data structure and algorithm choices.
- Measure the performance of an implementation and optimize it.
- Identify the operations being counted and resources being measured.
- Discuss the choices made in asymptotic complexity analysis (best, average, worst, amortized; space, time).
- Calculate the asymptotic complexity of an algorithm and algebraically manipulate standard notations for complexity.
- Classify and rank common algorithms and data structure operations by asymptotic complexity.
- Discuss memory as a resource.
- Trade off resource use, such as space/time.
- Discuss the details suppressed by asymptotic complexity analysis (e.g. performance measurement to reveal large constants).
- Discuss factors other than computational efficiency that influence the choice of algorithms, such as programming time, maintainability, use of application-specific patterns in the input data (per ACM), and ease of correctness/debugging.
- Classify algorithms by strategy, such as greedy, brute force, divide-and-conquer, decrease-and-conquer, transform-and-conquer, and dynamic programming.
Recognize data structures.
- Recall the data structures that are available to a program, by language, libraries, framework, and so forth.
- Recognize implicit structure, such as a heap implicit in an array, and the distinction between a list and a sorted list.
Implement data structures appropriately.
- Implement a dynamic container and the operations to manipulate it.
- Implement a static container with the optimizations that are possible because its shape is constant.
- Implement a container for mutable items, including any required type annotations and operations that adapt the structure when data change.
- Implement a container for immutable items with the optimizations that are possible because each item is constant.
- Distinguish the mutability of a reference from the mutability of the location it refers to.
- Write an algorithm using in-place, copying, and memory-adaptive strategies, as appropriate.
- Adapt an implementation to serve a different abstraction, such as a list presented as a stack, while being mindful of leaky abstractions (e.g. query and reserve the capacity of the array underlying a vector).
- Use nested data structures, such as arrays of lists and dictionaries of sets of strings.
- Write a data structure that presents a slice or view into another data structure.
- When copying or moving a data structure, consider how its items are duplicated, e.g. shallow by reference or deep by recursive copy, and choose appropriately.
- Implement an algorithm using inversion of control, as in map, reduce, visitors, or event callbacks, and choose between inversion and traditional control structures appropriately.