[Algorithms] Using Dynamic Programming to Solve longest common subsequence problem
Let's say we have two strings:
str1 = 'ACDEB'
str2 = 'AEBC'
We need to find the longest common subsequence, which in this case should be 'AEB'.
Using dynamic programming, we want to compare by char not by whole words.
- we need memo to keep tracking the result which have already been calculated
- memo is 2d array, in this case is 5 * 4 array.
- It devided problem into two parts
- If the char at the given indexs for both strings are the same, for example, 'A' for str1 & str2, then we consider
'A' + LSC(str1, str2, i1 + 1, i2 + 1)
-
- If the char at the given indexs are not the same, we pick max length between LCB('DEB', 'EBC') & LCB('CDEB', 'BC'), we pick
Max { LCS('DEB', 'EBC'), LCS('CDEB', 'BC') }
Bacislly for the str1 = 'CDEB' str2 = 'EBC', the first char is not the same, one is 'C', another is 'E', then we devide into tow cases and get the longer one. The way to devide is cutting 'C' from str1 get LCS('DEB', 'EBC'), and cutting 'E' from str2 get LCS('CDEB', 'BC').
/** * FIND THE LONGEST COMMON SEQUENCES BY USING DYNAMICE PROGRAMMING * * @params: * str1: string * str2: string * i1: number * i2: number * memo: array [] * * TC: O(L*M) << O(2^(L*M)) */ function LCS(str1, str2) { const memo = [...Array(str1.length)].map(e => Array(str2.length)); /** * @return longest common sequence string */ function helper(str1, str2, i1, i2, memo) { console.log(`str1, str2, ${i1}, ${i2}`); // if the input string is empty if (str1.length === i1 || str2.length === i2) { return ""; } // check the memo, whether it contians the value if (memo[i1][i2] !== undefined) { return memo[i1][i2]; } // if the first latter is the same // "A" + LCS(CDEB, EBC) if (str1[i1] === str2[i2]) { memo[i1][i2] = str1[i1] + helper(str1, str2, i1 + 1, i2 + 1, memo); return memo[i1][i2]; } // Max { "C" + LCS(DEB, EBC), "E" + LCB(CDEB, BC) } let result; const resultA = helper(str1, str2, i1 + 1, i2, memo); // L const resultB = helper(str1, str2, i1, i2 + 1, memo); // M if (resultA.length > resultB.length) { result = resultA; } else { result = resultB; } memo[i1][i2] = result; return result; } return { result: helper(str1, str2, 0, 0, memo), memo }; } //const str1 = "I am current working in Finland @Nordea", //str2 = "I am currently working in Finland at Nordea"; const str1 = "ACDEB", str2 = "GAEBC"; const { result, memo } = LCS(str1, str2); console.log( ` ${str1} ${str2} 's longest common sequence is "${result === "" ? "Empty!!!" : result}" ` ); console.log(memo);
----
Bottom up solution can be:
1. Init first row and first col value to zero
2. Then loop thought the data, If row latter and col latter is not the same, then take which is larger Max {the previous row same col value data[row-1][col], same row but previous col data[row][col-1]}
3. If they are the same, take data[row-1][col-1] + 1.
【推荐】国内首个AI IDE,深度理解中文开发场景,立即下载体验Trae
【推荐】编程新体验,更懂你的AI,立即体验豆包MarsCode编程助手
【推荐】抖音旗下AI助手豆包,你的智能百科全书,全免费不限次数
【推荐】轻量又高性能的 SSH 工具 IShell:AI 加持,快人一步
· SQL Server 2025 AI相关能力初探
· Linux系列:如何用 C#调用 C方法造成内存泄露
· AI与.NET技术实操系列(二):开始使用ML.NET
· 记一次.NET内存居高不下排查解决与启示
· 探究高空视频全景AR技术的实现原理
· 阿里最新开源QwQ-32B,效果媲美deepseek-r1满血版,部署成本又又又降低了!
· Manus重磅发布:全球首款通用AI代理技术深度解析与实战指南
· 开源Multi-agent AI智能体框架aevatar.ai,欢迎大家贡献代码
· 被坑几百块钱后,我竟然真的恢复了删除的微信聊天记录!
· AI技术革命,工作效率10个最佳AI工具
2015-01-01 [ES6] 16. Object Enhancements