QBUS3330 S2 An Excel file with Precision

AI 应用开发 × 算法面试题:业务场景实战 12 题【TopK、LRU、RRF 融合、相似度、限流算法,含手撕代码与底层答案】 本文聚焦AI应用岗算法面试新趋势,剖析12道融合业务场景的高频题:从海量日志中挖掘慢查询(分治+小顶堆)、亿级去重(布隆过滤器)、向量与关键词融合排序(RRF)、语义缓存设计(带租户隔离与TTL的LRU)到限流策略(令牌桶+滑动窗口),强调“场景复述→算法选型→手撕代码→生产落地”四步答题框架。突出考察点已从纯算法转向工程权衡、业务理解与系统设计能力,如误判容忍度、实时性与内存预算的平衡,助力开发者应对2026年真实技术挑战。 阅读详情


QBUS3330 S2 2024 Assignment 1
Page 1
This is an individual assignment. It is worth 10% of your final grade. It consists of three
questions. Each question is worth different marks. It is due on Friday 6 September at 11:59pm
and must be submitted through Canvas using Turnitin.
The submission will comprise two separate parts:
1. A typed report (PDF please) that addresses all questions and contains images all of
relevant tables, charts and decision trees within the report.
2. An Excel file with Precision Tree enabled containing all the original tables, charts and
decision trees. The report must be able to be read as a standalone document. The
Excel file is provided for backup and corroboration purposes.
Failure to submit both files by the due date will result in late penalties being applied.
Additional instructions occur after the questions.
Question 1 (25 marks)
Motivated by the looming Paris Olympics, back in 2023 the Australian Institute of Sport (AIS)
had to decide whether to introduce mandatory drug testing for athletes. Knowing that drug
tests are not completely reliable, they want to use decision tree analysis to see whether the
benefits outweigh the costs.
Probabilities:
If an athlete is tested for a particular drug, the test result will either be positive or negative.
However, as these tests are not completely accurate some athletes who are drug free test
positive (a false positive) and some athletes who are drug users test negative (a false
negative). The best data we have available suggests that 8% of all athletes use drugs. 2.5% of
all tests on DF athletes result in false positives and 7.5% of all tests on drug using athletes
result in false negatives.
Monetary Values:
The monetary values are difficult to assess but include the following:
 The benefit B from correctly identifying a drug user and banning them.
 The direct cost, C1, of a test.
 The cost, C2, of violating a non-users privacy by performing the test.
 The cost C3, of falsely accusing a non-user and banning them.
 The cost C4, of not identifying a drug user and allowing them to participate.
We measure the benefit and costs by indexing them against cost C1, which we assign a value
of minus one (-1). This does not mean the direct cost of testing an athlete is $1, it just means
that we express all other monetary values in multiples of C1. The index values are shown in
Table 1.1 on the following page.

QBUS3330 S2 2024 Assignment 1
Page 2
Table 1.1
Cost/benefit Index
C1 -1
C2 -2
C3 -20
C4 -10
B +25
Questions:
1. In Excel create two net benefit pay-off tables that map the net benefit of either testing
(four different states) or not testing (two different states) against the decision to ban
or not ban an athlete. Include the pay-off tables in your report. Note: the first table
should be expressed in index notation (+B, -C1, -C2 etc) while the second table should
state the net benefit in numerical terms based on values indicted in table 3.1. For
example, if a positive test is obtained for a non-drug user and this athlete is banned,
there are three associated costs: Cost of the test (-1), the cost of violating the athlete's
privacy (-2) and cost of falsely accusing the athlete (-50). (6 marks)
2. Calculate the relevant posterior probabilities. Include any Bayes tables generated in
Excel in your report. (4 marks)
3. Based on the values in the net benefit pay-off table and the Bayesian probabilities,
create a decision tree using Precision Tree that will help the AIS decide whether they
should implement mandatory drug testing. Note: Be careful to avoid double counting
the costs. Include your decision tree in your report. (8 marks)
4. For the given assumptions around the cost and benefit, outline the best strategy and
its net benefit and discuss this solution. (2 marks)
5. Conduct a brief sensitivity analysis giving reasons why you might change the relative
index values. Discuss how this might impact the original solution. (5 marks)

Question 2 (35 marks)
Intelligent Computing (iC) has been invited to submit by the University of Sydney  s
procurement office. The contract calls for the supply of 200 generic desktop computers and
associated accessories which will be used for digital in-place exams. All vendors must fulfil the
order within 6 weeks of contract award.
Despite the urgency the contract specifications are generic and so the university has informed
all bidders that the low bid will win the contract. iC believes that the cost of preparing the bid
will be $10,000 and the cost of supplying the computers will be $190,000.
QBUS3330 S2 2024 Assignment 1
Page 3
The bids are sealed, so iC has no information about the value of the bids their competitors
will submit. However, in the last 12 months iC has managed to poach several key employers
away from vendors who are competing for the contract and so iC has a good understanding
of how the competitors may behave. In summary iC believes that the size and probability of
a low competitor bid will be as follows:
Table 2.1
Low Bid Probability
Less than $230,000 0.20
Between $230,000 - $240,000 0.40
Between $240,000 - $250,000 0.30
More than $250,000 0.10
In addition, because of supply chain constraints, iC think there is a 30% chance there will be
no rival bid.
Part A (20 marks)
1. Based on this information create a pay-off table in excel which outlines iC  s most
logical bid prices against potential competitor bid options. Include relevant
probabilities. (6 marks)
2. Based on the pay-off table and any other relevant information use Precision Tree to
create a decision tree which sets out the problem. Include an image of the whole tree
in your report. (10 marks)
3. Using this decision tree, indicate the strategy that maximises EMV for iC. What is that
optimum value? (4 marks)
Part B (15 marks)
Use the sensitivity analysis function on the Precision Tree tool bar to vary the following inputs:
? Bid Preparation Costs: +/-10% in 1% increments
? Supply Cost: +/-10% in 1% increments
? No competing bid percentage: A minimum of 0% to a maximum of 60% in 5%
increments
1. Run a one-way sensitivity analysis on the entire decision tree model selecting one of
either the tornado graph or the spider graph. Include an image of the chosen graph in
your report and provide a short interpretation of the graph. (5 marks)
2. Run a second one-way sensitivity analysis on the bid value decision node (not the
entire tree) and create a strategy region graph. Include an image of the graph - that
plots the probability of no competing bids against Expected Value - in your report
along with a short interpretation of the graph. (5 marks)
QBUS3330 S2 2024 Assignment 1
Page 4
3. Run a two-way sensitivity analysis on the bid value decision node and create another
strategy region graph including an image of the graph in your report along with a short
interpretation. (5 marks)

Question 3 (40 marks)
The following table outlines the potential pay-offs for three separate investments.
Table 3.1
Investment A Investment B Investment C
Pay-off ($'000) Probability ($'000) Probability ($'000) Probability
1 18.0 10% 27.0 20% 18.0 20%
2 36.0 30% 45.0 30% 45.0 40%
3 61.0 30% 61.0 20% 72.0 20%
4 90.0 30% 99.0 30% 90.0 20%
Write a report for potential investors which ranks the three investments in terms of their
attractiveness. Your report should be approximately 500 words. All diagrams and tables used
to support your analysis should be generated in Excel using Precision Tree. As you do not
know the individual risk preferences of investors you should consider all risk types.

Other Instructions
Word count
1,000 words +/-10% excluding tables, decision trees and charts and references. Any words
beyond 1,100 will not be marked. Submissions below 900 words may be penalised.
Style
This is a business report not an essay. The report should:
- Have a suitable cover page.
- Be divided into 3 distinct sections.
- The text should be concise. Using bullet points is acceptable.
- Avoid the use of personal pronouns.
- Be professionally and logically laid out with good grammar and spelling.
- Marks will be deducted for submissions that do not meet these requirements.QBUS3330 S2 2024 Assignment 1
Page 5
Rubric
There general rubric relating to this assignment in the Assignment 1 Canvas module where
this brief is located that provides detailed information regarding the quality expectations of
the submission.
Late penalties
Reports submitted after the due date will incur a late penalty of 5% per day or part thereof.
Reports more than 10 days late will not receive a mark.
Academic integrity
This is an individual assignment. Collaboration is permitted to help understand the questions
otherwise all written work, research and analysis needs to be your own.
The university acknowledges that most students have access to generative artificial
intelligence (AI) technologies such as ChatGPT.
Students are permitted to use artificial intelligence (AI) tools such as ChatGPT etc. to generate
material that they will then evaluate, analyse and edit for incorporation into their reports.
Students must indicate when and where AI tools have been used. AI generated text cannot
be used or represented as original, self-authored work. Failure to declare the use of AI tools
will be reported as a major breach of policy for investigation.
Students are not permitted to use artificial intelligence (AI) tools such as ChatGPT etc. to
generate material that will be submitted unevaluated and unedited as part of their report. AI
tools cannot be relied upon to generate responses that will adequately address the
requirements of the assessment. Their undeclared use will be reported as a major breach of
policy for investigation.
In line with the Academic Integrity Procedures (2022), minor breaches of academic integrity
(including poor paraphrasing, failure to acknowledge the work of others or incorrect citations)
may be subject to a penalty of up to 15% of total available marks at the point of grading.
Major breaches of the Policy will be submitted for investigation to the Faculty's Academic

C++ 极致工程美学:拆解 Meta Faiss 量化器的 AVX2 向量化与有偏收缩算法 在 Meta 向量检索库 Faiss 的头文件 内部,标量量化器为反量化系数预留了两个并列枚举:无偏(UNBIASED)与有偏(BIASED)。按照经典数理统计教材的直觉惯性,无偏估计一向被奉为无损推断的金科玉律;但在 100 万条 768 维真实特征向量的检索压测中,把缩放模式切到有偏估计后,重构均方误差(MSE)稳定下降 19.60%,P99 长尾失真压低 35.05%,最大极端离群误差更是被削减了 81.46%。向量检索面对的是单次离散几何判决,没有任何大数定律能抹平单次重构方差。本文聚焦 Fais 阅读详情

相关推荐

vivado安装空间与依赖要求:入门须知要点

准备安装Vivado时,需重点关注系统空间和依赖环境的配置。结合vivado安装教程,详细说明最低磁盘要求、内存配置及操作系统兼容性,帮助新手快速搭建开发环境,避免常见安装失败问题。

weixin_34725745的博客 1581

【算法拾光】队列|在每个树⾏中找最⼤值 + 最后⼀块⽯头的重量

本文总结两道经典算法题:一是通过层序遍历(BFS)逐层找最大值,利用队列分层处理,每层维护最大值;二是用大根堆(优先队列)模拟石头碰撞过程,每次取最大两块石头,差值重新入堆,直至剩余一块或无石。前者适用于树的层次遍历,后者适合动态获取最值场景。核心思想:队列用于分层遍历,堆用于高效取极值。

wuqingyigexiaoba的博客 202

CANoe UDS $27服务DLL生成实战:从工程配置到算法修改全流程

本文详细介绍了在Vector CANoe中为UDS(统一诊断服务)的$27安全访问服务生成定制化DLL(动态链接库)的全流程。从定位官方参考工程、配置Visual Studio开发环境,到深入解析核心的GenerateKeyEx函数接口,并重点演示了如何修改和实现自定义的种子-密钥安全算法,最终完成DLL的编译、部署与CANoe集成测试,为汽车电子诊断开发提供实战指导。

star5的专栏 536

把 KV Cache 压到 FP4:BitDecoding-FP4 的算法-系统协同设计复盘

长上下文推理卡在KV Cache的显存与带宽上,而现有FP4方案要么精度崩、要么速度废。这篇文章复盘BitDecoding-FP4如何用三个算法补丁加一个系统重构,在B300上把服务吞吐提升2.4–3.0倍,精度几乎不掉。

helloagent的博客 319

逆向工程实战:从零解析微信PC版协议,并用开源代码构建你的私有化接口

本文详细解析了微信PC版的通信协议架构,从逆向工程基础到网络流量捕获、关键函数定位与HOOK技术实践,最终通过开源代码构建私有化接口。文章重点介绍了协议特征分析、HOOK技术应用以及安全实践,帮助开发者深入理解企业级通信协议的设计逻辑和安全机制。

weixin_30496999的博客 394

Miniforge3部署教程,完美替代Anaconda3

大背景不介绍了,哎!直接上教程!

zzy_ucas的博客 1万+

【算法拾光】链表 哈希表|K个一组翻转链表 + 两数之和

K个一组翻转链表需分段反转,借助哑节点与前驱指针维护连接,避免断链;两数之和利用哈希表实现O(1)查找,空间换时间,高效定位目标组合。前者考察链表指针操作,后者聚焦映射关系构建,均为面试高频题。

wuqingyigexiaoba的博客 341

与OpenClaw同类性质的AI智能体框架/工具汇总(小白友好)

OpenClaw的核心定位是「开源、本地优先、可执行真实任务的AI代理」,核心价值是让AI从“聊天”落地到“动手干活”(操控电脑/服务器、处理办公/系统任务)。结合这一核心,以下是同类性质的工具/框架,按「小白易上手程度」排序,每个产品重点说明和OpenClaw的异同、核心优势及适配场景,方便小白快速区分选择。

hy行者勇哥的博客 2111

排序算法深度笔记(下):非比较类排序的线性突破与工程边界

排序算法时间复杂度额外空间复杂度稳定性是否基于比较选择排序O(N^2)(O(1))不稳定是冒泡排序O(N^2)(O(1))稳定是插入排序O(N^2)(O(1))稳定是归并排序O(N\log N)(O(N))稳定是快速排序O(N\log N)O(log N)不稳定是堆排序O(N\log N)O(1)不稳定是计数排序(O(N + M))O(M))稳定否基数排序O(N)稳定否桶排序平均O(N) / 最坏O(N^2)

2401_87069538的博客 617

YOLOv8 骨骼结构识别详细说明

YOLOv8骨骼结构识别是一种基于深度学习技术的医学影像分析方法,主要用于X光、CT等影像中的骨骼定位与识别。该系统具有多骨骼同步识别能力,可应用于手部、脊柱、骨盆等不同部位,并能精确定位骨骼关键点。技术实现包括数据预处理、模型训练、推理检测、关键点分析和批量处理五个核心模块,支持骨折检测、骨龄评估等临床应用。该系统需注意医学合规性,要求高质量标注数据,并针对骨骼影像特点进行特征工程优化,同时考虑不同设备的部署需求。

u010117029的博客 334

SQL面试题——最大连续登陆问题

最大连续登陆是我们计算比较常用的指标,我们可以将其当成一个经典的SQL模型,后面我们也会陆续总结出其他的模型,你还知道其他的比较经典的SQL 吗?欢迎分享,或者你还知道其他的关于计算连续的指标?最大连续登陆重要的是计算思想,最大连续才是关键,登陆只是一个场景,要学会举一反三。

9014

优选算法:分治-归并排序

交易逆序对的总数

Wangshijie1015的博客 154

化工模拟与优化软件:Aspen HYSYS二次开发_(15).工业应用与项目管理

项目背景、目标和主要流程。

kkchenjj的博客 1131

408 算法 06|堆与 Top-K:最小的 10 个数,到底该用数组还是堆

2022(42) 找 100 万个数里最小的 10 个,官方给了"10 元素有序数组"和"大根堆"两种答案。我把两种都实现并数了比较次数:k=10 时两者几乎打平,k=10000 时堆少 111 倍。另附 2021(42) 计数排序的比较次数与稳定性实测。

m0_56226131的博客 453

SAP B1中文培訓教材_SDK

公司内部培训资料,对开发工具 ,DIAPI, UIAPI,Add-ON有详细解说

华为的C语言编程规范.pdf

华为技术有限公司内部技术规范。本规范制定了编写C语言程序的基本原则、规则和建议。从代码的清晰、简洁、可测试、安全、程序效率、可移植各个方面对C语言编程作出了具体指导。

算法64,字母异位词分组

代码核心思路是:互为字母异位词的字符串,在对其字符进行排序后,一定会得到相同的字符串。因此,可以将排序后的字符串作为哈希表的键(Key)来进行分组。

sheshibenniaodan的博客 219

C# WinForm 基于DirectShow.dll编写的程序,可边预览边保存视频,音频和图片

基于微软的DirectShow.dll的二次开发1.自动识别连接的摄像头跟麦克风且可以对其进行选择(默认选中第一个)2.可录制wmv和avi两种格式的视频并且都带有音频,avi可以录制的分辨率和帧数进行选择,wmv默认进行了淸晰度优化如不喜欢可去除代码中有相关提示3.可在预览点开之后进行拍照,也可边录制视频边拍照4.添加了视频和音频的压缩功能(酌情使用)这次二次开发发现C#对DirectShow的相关信息太少,这次做完也不敢藏私共享出来,还望大家指正

408 算法 04|主元素与三元组最小距离:两种“扫描时做决定“的思路

2013(41) 的主元素用摩尔投票一遍选出候选,但真正的考点是"选完必须验票";2020(41) 的三元组最小距离,三指针每次只挪最小的那个,靠的是 D=2×(max−min) 这个恒等式。两套完整 C 代码加真实输出,附官方评分档位表。

m0_56226131的博客 457

VC中sqlite的使用

在VC中对sqlite的使用,实现数据库大批数据的快速插入,读取数据库中的数据在ListCtrl中显示

牛客周赛165C题(前后缀性质,递推,排序,枚举)

题目大意:给定长度为 `n` 的数组,选择 `k ∈ [1,n]`,将前 `k` 个元素整体翻转,后面的元素不变。要求操作后的整个数组非递减,统计满足条件的 `k` 的数量。题目链接:https://ac.nowcoder.com/acm/contest/141721/C。时间复杂度:O(nlogn)

2501_94316951的博客 175

【MPC 算法】Python 实战:模型预测控制(MPC)算法原理与代码实现

本文介绍模型预测控制(MPC)原理与Python实现。MPC通过滚动优化预测未来状态,结合系统模型、目标函数与约束条件,动态生成最优控制序列。以一阶温度系统为例,使用CVXPY构建凸优化问题,实现加热功率的智能调控。代码包含模型预测、目标函数设计(含温度跟踪与控制平滑惩罚)、输入约束及滚动更新机制。仿真结果表明,MPC能快速稳定温度至目标值60℃,且控制动作平滑合理,体现其在复杂系统中的优越性。

赖德发的博客 70

算法61,判定是否为字符重排

两字符串能互为重排 ⇔。

sheshibenniaodan的博客 227

【前缀和+复合离散化】题解:双色矩形(two)_数据结构_算法优化_预处理_二维前缀和_STL_C++算法竞赛

给定一个 $n \times n$ 的黑白棋盘,初始按 $i+j$ 奇偶染色,经 $k$ 个格子颜色翻转后,求子矩形中黑白格子数差的绝对值最大值。由于 $n$ 可达 $10^9$,但 $k \leq 50$,可离散化关键点。枚举所有由修改点构成的矩形及其四边向外扩展(共 $16$ 种扩展方式),利用二维前缀和快速计算变化量,结合奇偶性判断多出的格子颜色,最终取最大绝对差。时间复杂度为 $O(k^4)$,在 $k \leq 50$ 下可行。

YLCHUP的博客 214

算法62&&63,存在重复元素

判断整数数组中是否有任意元素出现至少两次。:判断数组中是否存在两个。,且索引差值的绝对值。

sheshibenniaodan的博客 105

商业综合体公摊电费分摊算法:512 商户双层权重与尾差实测

公摊电费分摊算法是指把公共回路电量按公示权重拆到每户的规则化计算过程,核心不是权重多精巧,而是每张账单都能被租户手算复现。某珠三角商业综合体 512 户商户实测:固定费与变动费按 4:6 加权,尾差用余数递补压到 0.01 千瓦时精度,24 个月 12288 张账单重算差异为零,月度争议从 31 起降到 2 起。适用边界:商铺与租户已独立装表、公共回路已分项计量的商业综合体适用;末端未装表、靠总表减合计倒推公摊的场景不适用,先把表补齐再谈分摊。

185

考拉兹猜想 - 7

本文通过引入虚数单位重构Collatz猜想的迭代逻辑,指出偶数减半与奇数“乘3加1”的过程在虚数坐标下等价于归结操作。关键在于:乘3加1使奇数跃升至偶数,随后持续减半,而虚数单位代表“无限减1”,实现无限向有限的转化。无论初始值如何,减半始终主导,最终归结为4-2-1循环。该视角消解了正向思维中的速度竞争疑虑,证明归结是必然结果。

杨铸人的博客 93

2026华为OD机试真题 10月14日 新系统 【分组字符串排序】

华为OD机试真题 新系统【分组字符串排序】,提供C语言、Java、C++、Python、Go、JavaScript源码实现,并提供完整思路讲解,以及在线OJ模拟做题。华为OD机试真题 【分组字符串排序】多语言题解。

qq_45776114的博客 459

贪心题目:可以到达的最远建筑

为了到达尽可能远的建筑物,在将砖块替换成梯子时,应将高度差最大位置的砖块替换成梯子,此时剩余的砖块数量最多,如果将高度差较小位置的砖块替换成梯子,则剩余的砖块数量较少,能到达的最远建筑物下标不可能更大。开始向后面的建筑物移动,移动过程中维护优先队列存储正高度差,优先队列的队首元素是最大元素,每次遇到正高度差时优先考虑使用砖块,如果砖块数量不够则使用梯子替换最大高度差位置的砖块,直到砖块和梯子都使用完毕时,到达的建筑物即为最远建筑物。如果以最佳方式使用给定的梯子和砖块,返回可以到达的最远建筑物的下标。

stormsunshine的博客 301

强化学习第十一课:完整 Q-Learning 算法

表示有:\[ 5 \]0 1 2 3 40:向左1:向右目标位置:\[ 4 \]

qq_56657939的博客 328
上一篇: Data Visualisation for Managers (INFS6023)
下一篇: ECON0013: MICROECONOMICS
vx_codinghelp
博客等级 码龄2年 543粉丝 68原创
评论
成就一亿技术人!
拼手气红包6.0元
还能输入1000个字符
 
 条评论被折叠 查看
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付:10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值