题意：

在一个2D游戏背景下设计和分析算法以提高对算法和递归关系复杂度的理解

解析：

游戏背景：

M*M大小的笛卡尔坐标系网格，蓝色点代表AI控制的敌人，红色叉代表玩家，虚线框代表玩家的攻击范围，是一个边长 b =1.5 的正方形。

（1）：完成一个O（1）的函数，判断点（x,y）是否在玩家攻击范围之内——
function IsAffected(($p_x, p_y$px,py),($x, y$x,y), $b$b)
根据提供的玩家坐标($p_x,p_y$px,py)计算出攻击范围的边界，x方向上（$p_x- b，p_x+ b$px−b，px+b），y方向上（$p_y- b，p_y+ b$py−b，py+b），判断坐标即可

（11）：判断敌人是否通存在通信，本质上就是判断两个敌人的连线是否穿过攻击区
function CanDirectlyCommunicate($x_1$x1，$y_1$y1),（$x_2$x2，$y_2$y2）,($p_x$px，$p_y$py),$b$b)
敌人两点确定一条直线，就是他们的通信路线，攻击区域范围为一个正方形($x_1+b&lt;x&lt;x_2+b,y_1+b&lt;y&lt;y_2+b$x1+b<x<x2+b,y1+b<y<y2+b)(假设$x_1&lt;x_2,y_1&lt;y_2$x1<x2,y1<y2)，通过数学关系来判断直线是否经过正方形区域

涉及知识点
分治与递归，算法时间复杂度分析，邻接矩阵

更多可加微信讨论

微信号：ITCSdaixie

pdf全文
The University of Melbourne
School of Computing and Information Systems
COMP90038 Algorithms and Complexity
Assignment 1, Semester 2 2019
Released: Friday August 23 2019. Deadline: Sunday September 8 2019 23:59
This assignment is marked out of 30 and is worth 15% of your grade for COMP90038.
Objectives
To improve your understanding of the time complexity of algorithms and recurrence relations. To develop problem-solving and design skills. To improve written communication skills; in particular the ability to present algorithms clearly, precisely and unambiguously.
Problems
Let’s play a game. Or, more precisely, design and analyse some algorithms that might be used in a simple two-dimensional (2D) game.

Consider a game played on an two-dimensional grid, whose Cartesian coordinates range from (−M, M). . .(M, M). Figure 1 depicts a game board for which M = 4.
The game contains a fixed number N of enemy players, who are AI-controlled. Each player (including the human player and each enemy AI) is located at some arbitrary but fixed position (x, y) on the board,i.e. for which −M ≤ x ≤ M and −M ≤ y ≤ M.
The human player can perform certain attacks, which damage all enemy players within a certain range of the human player. To avoid the need to calculate with expensive multiplication and square root operations, we will approximate the range by a square situated about the human player.
Specifically, given some fixed bound b, for which 0 ≤ b ≤ M, if the human player is located at position ($p_x$px, $p_y$py) then all enemy players that lie within or on the borders of the box whose bottom-left corner is ($p_x−b$px−b, $p_y−b$py−b) and whose top-right corner is ($p_x+b$px+b, $p_y+b$py+b) are affected by the attack. Figure 1 depicts an example box for the bound b = 1.5.

1. [1 mark] Implement an Θ(1) function to determine whether an enemy at position (x, y) is affected by the player’s attack, given the player’s location ($p_x$px, $p_y$py):
   function IsAffected(($p_x$px, $p_y$py),($x$x, $y$y), $b$b)
   . . .

2. [2 marks] Suppose (for now) that the positions of the enemy players are stored in an unsorted array A[0] . . . A[N − 1].
   The following function uses the divide and conquer approach to mark all enemy players affected by a player attack. Marking is implemented by the Mark function whose details are not important. (Note: the division below is integer division, i.e. it rounds down to the nearest integer.)
   function MarkAffected(($p_x$px, $p_y$py), A, b)
   MarkAffectedDC(($p_x, p_y$px,py), A, 0, N − 1, b)
   function MarkAffectedDC(($p_x, p_y$px,py), A, lo, hi, b) . assume lo ≤ hi
   if lo = hi then
   if IsAffected((px, py), A[lo], b) then
   Mark(A[lo])
   else
   mid ← lo + (hi − lo)/2
   MarkAffectedDC(($p_x, p_y$px,py), A, lo, mid, b)
   MarkAffectedDC(($p_x, p_y$px,py), A, mid + 1, hi, b)
   Explain the purpose of the outermost “if/else” test. In particular, suppose we removed the line “if lo = hi” and the “else” line. In no more than a paragraph explain how this would affect the algorithm

3. [2 marks] Consider the following recurrence relations. Which one best describes the worst-base complexity of the MarkAffectedDC function, whose input size is n = hi −lo + 1 and whose basic operations are calling IsAffected and Mark? Justify your answer in no more than a paragraph of text.
   T(1) = 1
   T(n) = 2T(n/2)
   T(1) = 2
   T(n) = 2T(n/2) + 2
   T(1) = 0
   T(n) = 2T(n/2)
   T(1) = 2
   T(n) = 2T(n/2)

4. [2 marks] Use telescoping (aka back substitution) to derive a closed form for your chosen recurrence relation and thereby prove that the worst case complexity of the MarkAffectedDC algorithm is in Θ(n).

5. [2 marks] Can we do better than this? Recall that the human player is at some fixed location ($p_x, p_y$px,py). Your task is to work out how you would sort the array A so that those enemy AIs that need to be marked can be identified in log(N) time.
   Specifically, complete the following comparison function that would be used while sorting the array A. Here, ($x_1, y_1$x1,y1) and ($x_2, y_2$x2,y2) are two points from the array A. The function should return true if it considers the first point to be less than or equal to the second, and should return false otherwise. Your function can use the player’s coordinate (px, py) as global variables, i.e. you are allowed to refer to px and py in your function.
   function LessOrEqualTo(($x_1, y_1$x1,y1),($x_2, y_2$x2,y2))
   . . .

6. . [3 marks] Now, supposing the array has been sorted using your comparison function, implement an algorithm whose worst case complexity is in Θ(log(N)) that determines which array elements should be marked. Your function should take the bound $b$b as an argument, and may also take the player’s coorinates ($p_x, p_y$px,py).

7. . [2 marks] How many elements will need to be marked in the worst case and what, therefore, would be the worst-case complexity of an algorithm that, given an array A sorted according to your comparison function, implemented the behaviour of MarkAffected above? Express your answer in Big-Theta notation.

8. . [2 marks] The worst case is quite pessimistic. On average, we might expect that enemy AIs are uniformly distributed throughout the game board.Let d be the expected number of enemy AIs contained in or on the borders of a box of bound b(i.e. whose width and height are each 2b) on the board. For simplicity, we limit our attention to boxes that are contained entirely within the game baord.Consider an algorithm that, given an array A sorted according to your comparison function,implemented the behaviour of MarkAffected above by first using your Θ(log(N)) function to determine the elements that need to be marked (of which we expect there to be d), and then marking those d elements. We might characterise its expected complexity as:
   log(N) + d
   Derive a formula for d in terms of b, M and N.

9. [2 marks] As a point of comparison, what is the expected complexity of the divide-and-conquer
   algorithm above in this average case, expressed in Big-Theta notation? Justify your answer in no
   more than a paragraph of text.

10. [2 marks] Now compare the best case complexity of the divide-and-conquer algorithm and the one
    that uses a pre-sorted array A described above. Express the best-case complexity of each in BigTheta notation as a function of N. Justify both in no more than a paragraph of text.

11. [4 marks] Suppose now that the enemy AIs can send each other messages. However, two AIs,
    whose positions are (x1, y1) and (x2, y2) respectively, can directly communicate only if the line
    (x1, y1) — (x2, y2) that connects them does not pass through the rectangle (including its borders)
    surrounding the human player whose bottom left corner is (px−b, py −b) and whose top-right corner
    is (px + b, py + b), where (px, py) is the human player’s position.
    Implement a function that determines whether two enemy AIs can directly communicate. Include
    a brief description (no more than a single paragraph) of how your function works, i.e. the rationale
    behind its design.
    function CanDirectlyCommunicate((x1, y1),(x2, y2),(px, py), b)
    . . .

12. [6 marks] Imagine that your CanDirectlyCommunicate function has been used to construct a
    graph whose nodes are the enemy AIs, such that there exists an edge between node n1 and n2 if and
    only if those two AIs can directly communicate. This graph is stored as an adjacency matrix M.Here the nodes ni are simply the indices of enemy AIs in the array A, so each of them is simply an integer in the range 0 ≤ ni ≤ N −1. Therefore the adjacency matrix M is simply a two-dimensional array, M[0][0] . . . M[N − 1][N − 1].
    Implement an O(N2) algorithm that, given two enemy AIs n1 and n2 determines whether there exists a path by which they might communicate via other enemy AIs. If a path exists, your algorithm should return the shortest path by which communication is possible. Otherwise, your algorithm should return the empty path. Your algorithm also takes as its input the adjacency matrix representation of the graph described above.
    You should represent a path as a linked list of enemy AI indices ni. The empty path is therefore the empty linked list without any nodes.

Submission and Evaluation
• You must submit a PDF document via the LMS. Note: handwritten, scanned images, and/or Microsoft Word submissions are not acceptable — if you use Word, create a PDF version for submission.
• Marks are primarily allocated for correctness, but elegance of algorithms and how clearly you communicate your thinking will also be taken into account. Where indicated, the complexity of algorithms also matters.
• We expect your work to be neat – parts of your submission that are difficult to read or decipher will be deemed incorrect. Make sure that you have enough time towards the end of the assignment to present your solutions carefully. Time you put in early will usually turn out to be more productive than a last-minute effort.
• You are reminded that your submission for this assignment is to be your own individual work. For many students, discussions with friends will form a natural part of the undertaking of the assignment work. However, it is still an individual task. You should not share your answers (even draft solutions) with other students. Do not post solutions (or even partial solutions) on social
media or the discussion board. It is University policy that cheating by students in any form is not permitted, and that work submitted for assessment purposes must be the independent work of the student concerned.
Please see https://academicintegrity.unimelb.edu.au

If you have any questions, you are welcome to post them on the LMS discussion board so long as you do not reveal details about your own solutions. You can also email the Head Tutor, Lianglu Pan ([email protected]) or the Lecturer, Toby Murray ([email protected]). In your message, make sure you include COMP90038 in the subject line. In the body of your message, include a precise description of the problem.

Late Submission and Extension
Late submission will be possible, but a late submission penalty will apply: a flagfall of 2 marks, and then 1 mark per 12 hours late.
Extensions will only be awarded in extreme/emergency cases, assuming appropriate documentation is provided simply submitting a medical certificate on the due date will not result in an extension.