Effective resource allocation involves setting resource requests and limits for your pods. This ensures that Kubernetes schedules your containers efficiently without over-provisioning resources. Properly defining these parameters helps in maintaining a balanced load across the cluster and prevents resource contention.

Horizontal Pod Autoscaling (HPA) is crucial for optimizing Kubernetes Engine on GCP as it automatically adjusts the number of pod replicas based on observed metrics such as CPU and memory utilization. This ensures that your application can handle varying loads dynamically, maintaining optimal performance and resource usage.

Enabling Cluster Autoscaler helps in optimizing Kubernetes Engine on GCP by automatically adjusting the size of your cluster based on resource demand. It adds nodes when there is an increase in workload and removes them when they are no longer needed, ensuring efficient use of resources and cost savings.

Cost optimization strategies for Kubernetes Engine on GCP include using preemptible VMs for non-critical workloads, leveraging node auto-provisioning, and taking advantage of sustained use discounts. Additionally, rightsizing nodes and pods based on actual resource requirements can significantly reduce costs while maintaining performance.