What is the Problem with Logistics?
- Spatial Data – These are the data that contains longitude, latitude and time-stamp which makes it extremely difficult to mine without using a sophisticated programming language.
- No anomalies tracking system – Currently there is no way to detect if there are any anomalies in the performance of 3PL and this can lead to serious performance issues in Logistic department. This can cause high variability in lead-time and several problems related to warehouse and logistics.
- Excess Pipeline Inventory – Inefficient Logistics can lead to increase of pipeline inventory which can further lead to increase in production time
- No power to predict when there will be delivery – Currently there is now way to estimate when certain material will be delivered to the location, this may lead to inefficient supply chain and misunderstandings between various players in the chain.
- Reverser Logistic – Sometimes there are product returns and there is no way to predict this event and prepare in advance.
- Long route – Most of the time trucks and other freight service takes longer routes which results in longer delivery time and unsatisfied customer
How Machine Learning can help making Logistic Smarter
- Natural Language Processing/Text Mining – – About 90% of the data are in unstructured or text format. Using NL and Text Mining we can extract relevant information from this data for actionable business insights.
- Predictive Analytics – Predictive Analytics in Logistics refers to the use of available spatial, and warehouse data in order to perform analysis that determines patterns and predicts future outcomes and trends. Predictive analysis is comparatively a new term which has been introduced in procurement.
- Perspective Analytics: – Prescriptive analytics uses optimization or embedded decision logic rules to find out what should be done in a certain situation. This form of analytics is the most advanced of existing analytics, as it can provide the best way forward given specific business variables, inputs and objective.
- Reinforcement Learning – Reinforcement learning can be thought of as an agent interacting with an environment. The agent performs actions and receives rewards and builds up a picture of the environment (and how its actions effect the environment) through observations.
- Deep Learning – Deep Learning is a subfield of machine learning concerned with algorithms inspired by the structure and function of the brain called artificial neural networks
- Genetic Algorithm – Genetic algorithms are randomized search algorithms that have been developed in an effort to imitate the mechanics of natural selection and natural genetics. GA is stochastic search algorithm.
- Optimization Algorithm – This includes linear programing, integer programing, queuing theory, inventory optimization, transportation problem algorithm and transshipment problem algorithm
- Ranking Algorithm – Ranking algorithms determine the relative importance of objects in connection with other objects in the dataset. The page rank algorithm is probably the most well-known example as it’s extensively used by Google on their search engine results page. There are other algorithms for Ranking where research in going on. Using ranking algorithm we can rank our 3PL according to their performance and this also led to better performance Judgement.