CSE175-Assignment 1 Solved

This programming assignment has three main learning goals. First, the assignment will provide you with an opportunity to practice your skills developing Python programs in the PyCharm integrated development environment (IDE). Second, this assignment will provide you with some experience in implementing basic heuristic search algorithms, including greedy best-first search and A* search. For comparison, you will also implement the uninformed uniform-cost search algorithm. Third, this assignment requires you to design an admissible heuristic function in the domain of searching for a shortest path on a map. This will provide you with some experience exploring the features that make for good heuristic functions. A foundational understanding of these basic approaches to heuristic search will support your learning of more advanced techniques later in this class.

In summary, you will implement the uniform-cost search, greedy best-first search, and A* search algorithms in Python in the context of a simple shortest-path map search problem. These implementations will also support the optional checking of repeated states during search. You will also design an admissible heuristic function, with the goal of demonstrating a substantial improvement in performance over uninformed search methods.

Activities
You are to providePython functionsthat implementthe followingthree search algorithms: uniformcost search, greedy best-first search, and A* search. Your provided Python source code must be compatible with provided Python utility code which implements simple road maps, allowing your search algorithms to be used to find the shortest routes between locations on such maps. Indeed, your assignment solution will be evaluated by combining your submitted files with copies of the provided utility files and testing the resulting complete program against a variety of test cases. In other words, your solution must work with the provided utilities, without any modifications to these provided files.

More specifically, you are to provide the following functions in the corresponding files, implementing the corresponding algorithms:

Function
File
Algorithm
uniformcostsearch
“ucost.py”
uniform-cost search
greedysearch
“greedy.py”
greedy best-first search
astarsearch
“astar.py”
A* search
The source code for each of these functions should be very similar to that for the others. Rough pseudocode for these functions has been provided during class meetings. These functions must have the following features ...

•     takes two or three arguments:

1.    problem — a RouteProblem object

2.    h — a HeuristicFunction object (not included in uniformcostsearch)

3.    repeatcheck — a boolean indicating if repeated state checking is to be done

•     implements the pseudocode for generic search provided during class lectures

•     deviates from this pseudocode only by performing repeated state checking if and only if the repeatcheck argument is True

•     in particular, the goal test is performed on a search tree node just before it is expanded • makes use of a Frontier object to maintain the search tree fringe

•     returns a search tree node corresponding to a solution, or None if no solution is found

In general, your functions should allow the provided “main.py” script to output correct solutions (including path cost and expansion count statistics) for any map search problem provided as input to it. Note that this means that the functions that you implement should write no output, as this will clutter the output produced by the “main.py” script. If you include any statements that write output in your code (perhaps as tools for debugging) these should be removed prior to submitting your code files for evaluation. You may receive no credit for your submitted solution if it produces extraneous output.

In addition to these search algorithm functions, you must also provide a definition for a class called HeuristicFunction, in a file named “heuristic.py”, that implements an admissible heuristic function that can be quickly calculated. You are required to design this heuristic function by yourself, and the quality of your heuristic function will have a substantial influence on how your submitted assignment is evaluated. Your heuristic function must absolutely be admissible, but it should otherwise reflect as accurate an estimate of the residual path cost from a given search tree node as possible, given the constraint of rapid calculation. Note that locations have longitude and latitude coordinates that may assist in this process. Also note that, for the purpose of this assignment, the cost assigned to road segments is to be taken as a time cost. In other words, the goal of the search is to find the shortest path in terms of travel time, and each road segment is labeled with the time it takes to traverse that segment. This means that any measure of physical distance will not suffice as an admissible heuristic function, as such measures do not reflect an estimate of the remaining travel time to the destination. If you cannot think of a solution that is admissible regardless of the units of the map measurements, you may assume that location coordinates are measured in miles from an origin point and road segment costs are expressed in minutes. If you make an assumption about the units of measurement, or any other assumptions, you should indicate this fact in a comment in your submitted “heuristic.py” file. Keep in mind that your search algorithms should work for any valid map search problems provided as input to them.

The Python utility code that you are required to use is provided in a ZIP archive file called “PA1.zip” which is available in the “Assignments” section of the class CatCourses site, under “Programming Assignment #1”. These utilities include:

•     RoadMap class — This class encodes a simple road map, involving locations connected by road segments. Each location has a name and coordinates, which can be conceived as longitude and latitude. Each road segment represents a one-way connection from one location to another. Each road segment has a name and a cost of traversal.

•     RouteProblem class — This class encapsulates a formal specification of a search problem. It includes a RoadMap, as well as a starting location and goal location. It also provides a goal test function called isgoal.

•     Node class — This class implements a node in a search tree. Each node has a corresponding location, a parent node, and a road segment used to get from the location of the parent to the location of the node. (Note that the root of the search tree, corresponding to the starting location, has no parent or road segment.) Each node also tracks its own depth in the search tree, as well as the partial path cost from the root of the treee to the node. Each node also records the value of the heuristic evaluation function applied to the node’s location. The class provides an expand function, which expands the node.

•     Frontier class — This class implements the frontier, or fringe, of a search tree. The root node of a search tree is provided upon creation, initially populating the frontier with that one node. These objects are implemented as priority queues, releasing nodes in order of increasing values of some measure. At the time of the creation, the measure to be used to sort the nodes in the frontier must be specified: ’g’ (partial path cost), ’h’ (heuristic evaluation function value), or ’f’ (the sum of the other two measures). This class provides functions to add a node to the frontier and pop a node from the frontier, as well as testing if the frontier isempty or if it contains a node with a matching location.

The contents of these utility files will be discussed during a laboratory session, and comments in these files should assist in your understanding of the provided code. Questions are welcome, however, and should be directed to the teaching team.

Your implementations of all three search algorithms should largely mirror the generic search algorithm presented during class lectures. Specifically, your code must test for goal attainment just prior to expanding a node (not just prior to insertion into the frontier). No repeated state checking should be done unless the repeatcheck argument is True. When repeated state checking is being done, a child node should only be discarded if its state matches that of a previously seen node. (To be clear, a child node with a state that matches that of a node currently in the frontier may or may not be discarded due to repeated state checking, depending on the specific algorithm being implmented and the relative “costs” of the two nodes.) Note that the algorithm presented in class will require a slight modification to allow for the disabling of repeated state checking, based on the boolean argument provided to the search function. In general, your implementations should not depend on recursion to traverse the search tree, and they should make explicit use of a Frontier object to keep track of the fringe.

In order to obtain some confidence that your search algorithms work for any valid map search problems provided as input to them, as required, it is very likely that you will have to test your solution on a variety of test cases. A simple test case appears in the “main.py” script, but this test case is insufficient to fully test your code