Synthetic Datasets on the Product Portfolio Planning (PPP) Problem
This repository contains 135 synthetic datasets of different sizes for the product portfolio planning (PPP) problem. The synthetic datasets were first used to verify a novel probability-based discrete particle swarm optimization (PSO) algorithm for the PPP problem. Please refer to the following paper for more information about the PSO algorithm and the PPP problem datasets: Liu X, Li A. An improved probability-based discrete particle swarm optimization algorithm for solving the product portfolio planning problem. Research Square; 2022. DOI: 10.21203/rs.3.rs-1860671/v1. Each dataset is contained in a folder named “AXX-LXX-SXX-CXX-RXX”. “A” denotes the number of attributes (including price), “L” denotes the number of levels of each attribute, “S” denotes the number of segments, and “C” denotes the number of competitive products in the market. For a “AXX-LXX-SXX-CXX”, 5 datasets are generated. “R” is used as ID for datasets with the same “AXX-LXX-SXX-CXX”. In each folder, two CSV files named “probinfo.csv” and “data.csv” are included. The descriptions of the two files are shown below. 1. probinfo.csv This file provides related information for the dataset. LSL: The lower specification limit in the process capability index (PCI) (can be changed). beta: A constant (the average cost per variation of process capabilities) (can be changed). maxProdNum: The maximum number of potential products (can be changed). (In the paper, it is set as J=5 or J=8.) muP: A scale parameter in the multinomial logit (MNL) choice model (can be changed). AttrNum: The number of attributes. numLevels: The number of levels of each attribute. Prices: possible prices for a product (can be changed, but the number of prices should be equal to numLevels). numSeg: The number of segments. SegSize: The size of each segment. Competitive Product Num: The number of competitive products. Product 1: The setting of the first competitive product. Product 2: The setting of the second competitive product. AttrIndex: Just Used to generate the dataset, can be neglected during using the dataset. Other parameters used for solving the PPP problem should be given by researchers in their programs. 2. data.csv The part-worth utilities and standard times are stored in this file. AttribeteLevel: ID to show different attribute levels. A1-1 means the first level for the first attribute. s1-s4: Part-worth utilities for each attribute level for each segment (segments 1 to 4). time: Mean value of the part-worth standard time for each attribute level variance_time: Standard deviation of the part-worth standard time for each attribute level utility_weight: Weight for each attribute level, which is used to calculate the overall utility for a product. cost_weight: Weight for the time of each attribute level, which is used to estimate the overall cost for a product.
National Natural Science Foundation of China